From the SelectedWorks of Julian Dibbell
April 4, 2015
Invisible Labor, Invisible Play: Online Gold
Farming and the Boundary Between Jobs and
Games
Julian Dibbell
Available at: http://works.bepress.com/julian_dibbell/1/
INVISIBLE LABOR, INVISIBLE PLAY:
ONLINE GOLD FARMING AND THE BOUNDARY BETWEEN JOBS AND
GAMES
When does work become play, and play work? Courts have considered the question in a variety of
economic contexts, from student athletes seeking recognition as employees to professional blackjack players
seeking to be treated by casinos just like casual players. Here I apply the question to a relatively novel
context: that of online gold farming, a gray-market industry in which wage-earning workers, largely based in
China, are paid to play online fantasy games (MMOs) that reward them with virtual items their employers
sell for profit to the same games’ casual players. Gold farming is clearly a job (and under the terms of service
of most MMOs, clearly prohibited), yet as I show, U.S. law itself provides no clear means of distinguishing
the efforts of the gold farmer from those of the casual player. Viewed through the lens of U.S. labor and
employment law, the unpaid players of a typical MMO can arguably be classified as employees of the
company that markets the game. Viewed through caselaw governing when the work of professional players
does and does not constitute game play, gold farmers arguably are players in good standing. As a practical
matter, these arguments suggest new ways of approaching the regulation of so-called virtual property and of
online gaming in general. More broadly, the very viability of these arguments shows that the line between work
and play is not so much an empirical fact as it is a social one, produced by negotiations in which the law has
a leading role to play. This insight contributes to an ongoing debate about commodification and play that
grows more urgent as digital technologies suffuse the world’s economy with gaming and its logic.
INTRODUCTION ........................................................................................................ 2
I. GOLD FARMING: A CULTURAL, ECONOMIC, AND LEGAL OVERVIEW......... 5
A. Farming as Play: Gaming Culture in MMOs ............................................. 5
B. Farming as Production: The Economies of MMOs ............................... 10
C. Farming as Problem: Gold Farming and the Law .................................. 15
II. COMMODIFICATION AND PLAY ...................................................................... 22
A. The Commodification Debate................................................................... 22
B. The Commodification of Play ................................................................... 27
1. Play theories and commodification ....................................................... 27
2. Commodified play and the law .............................................................. 33
III. ARE UNPAID FARMERS EMPLOYEES? .......................................................... 38
A. Do Players Meet the Control Test of Employment? ............................. 40
B. Do Players Meet the Economic Test of Employment? ......................... 45
C. What Happens if Unpaid Farmers Are Employees? .............................. 53
IV. ARE PAID FARMERS PLAYERS? ...................................................................... 58
A. Are Gold Farmers Actually Playing the Game? ...................................... 58
B. What Happens if Paid Farmers Are Players (and if They Are Not)? ... 61
CONCLUSION ........................................................................................................... 64
DRAFT; please do not cite or circulate without author’s express permission
INTRODUCTION
By the best available estimates, roughly 400,000 people are employed
as gold farmers, all of them in developing countries outside the U.S., the vast
majority of them in China.1 The job of a gold farmer is to play a single online
computer game all day, all week. The game is always of a genre known as
massively multiplayer online role-playing games, or MMOs. Typically the
game is World of Warcraft, the most popular of the MMOs, with 7 million
paying players worldwide, down from 11 million at its height but still the
most popular in the class. The gold farmer does things all the other players
do: Kills monsters; collects coins, weapons, and other loot from the corpses
of the monsters; kills more monsters; collects more loot; takes loot into town
and sells it to other players for more coins. This routine is known as farming,
or grinding. All players must deal with it if they want to advance in the game.
But there is a great variety of other, often more interesting activities, and few
players pursue the grind as relentlessly as gold farmers. The few who do are
known as power gamers, and while their style of play, in particular, is hard to
distinguish from a gold farmer’s, their incentives for playing, like those of all
1
Richard Heeks, Virtual Economies, Virtual Goods and Service Delivery in Virtual Worlds,
2 J. VIRTUAL WORLDS RESEARCH 1, 7 (2010). All remaining facts about gold
farmers and online games are from Julian Dibbell, The Life of the Chinese Gold Farmer,
THE NEW YORK TIMES MAGAZINE, June 17, 2007,
http://www.nytimes.com/2007/06/17/magazine/17lootfarmers-t.html.
Page 2 of 68
DRAFT; please do not cite or circulate without author’s express permission
the other players, differ from the gold farmer’s in one important respect: The
gold farmer is paid a wage, in real money, to play the game.
Generally, in fact, a Chinese gold farmer’s playing environment is less
like a typical Chinese gamer’s than like a typical Chinese factory worker’s. He
plays in a workshop surrounded by other gold farmers. He works twelve
hour shifts and seven-day weeks and sleeps in a dorm with the other gold
farmers he works with. He punches a clock at the start of his shift, and at the
end of his shift he hands over the product of his work day, a modest quantity
of virtual gold coins, to the factory owners who pay his wage. The owners
then sell those coins, for real money, to an online retailer who will sell them,
for more real money, to players in the U.S. or Europe who have less time or
patience for the grind than most.
There is a mildly illicit quality to gold farming work. Like most MMO
companies, the one that runs World of Warcraft, Blizzard, Inc., bans the sale
of virtual items for real currency (a practice known as real-money trading or
RMT), and gold farm operators work hard to avoid detection by Blizzard lest
they lose their accounts and inventory. But the work of gold farmers is hardly
invisible—at least not in the way that theorists of labor have come to mean
when they apply that term to the work of domestic caregivers, prisoners, and
others excluded from conventional definitions of employment. If gold
farming took place on U.S. soil, for example, it would fall squarely within the
terms of the Fair Labor Standards Act2, the National Labor Relations Act3,
and most other elements of the modern labor-regulatory regime. Aside from
2
3
Fair Labor Standards Act (FLSA), 29 U.S.C. §§ 201–219.
National Labor Relations Act (NLRA), 29 U.S.C. §§ 151-169.
Page 3 of 68
DRAFT; please do not cite or circulate without author’s express permission
the intangibility of the goods gold farmers produce, after all, very little
distinguishes their working conditions from those of typical low-wage
manufacturing employment.
Yet there is another, less recognizable sort of labor hidden in gold
farming’s shadow: the farming that is done, as most MMO farming is, for no
monetary compensation at all. Convention requires that we call this unpaid
effort play, but given its stark similarities to what gold farmers do all day, it is
worth asking why, exactly, we should not call it a job. And if that question
bears asking, then so does another: Why not call what the gold farmers do a
game?
This article explores the implications of these two interlocking
questions both for the regulation of online games and, more broadly, for the
contemporary relationship between work and play. Part I describes the
relevant aspects of MMO farming, as experienced both by ordinary gamers
and by gold farmers. It also discusses the handful of legal cases in which gold
farming has been explicitly at issue. Part II locates gold farming, as an
instance of commodified play, within long-running debates about the
commodification of “priceless” goods in general and of play in particular.
Part III weighs the plausibility of legally classifying unpaid farming as
employment, and sketches the consequences that might follow from doing so.
Part IV, conversely, considers whether, as a legal matter, gold farmers are
actually playing the games in which they work, and how it might matter if
they are. The Conclusion points toward broader implications, arguing that
gold farming’s blurring of the lines between work and play reflects both the
Page 4 of 68
DRAFT; please do not cite or circulate without author’s express permission
destabilizing effects of networked technology on existing concepts of labor
and employment and the inherent instability of the concepts themselves.
I. GOLD FARMING: A CULTURAL, ECONOMIC, AND LEGAL OVERVIEW
A. Farming as Play: Gaming Culture in MMOs
To understand farming and its place in the culture of MMOs, it is
important to understand what MMOs are and what they are not. MMOs are
virtual worlds—online simulated environments in which users, embodied in
3D graphical representations known as avatars, interact both with their
simulated surroundings and with each other. But they are not the only type
of virtual world. So-called “sandbox” or “open” virtual worlds like Second
Life, for example, seek to mirror the real world’s range of open-ended
possibilities, offering users an almost limitless freedom to build their own
objects and invent their own goals. An MMO, on the other hand, is by
design a far more constrained experience. It is above all a game, with a rich
but bounded set of pre-defined goals and rules for attaining them.
More specifically, it is a variant of what the digital-games industry
refers to as a role-playing game, or RPG. In a typical single-player RPG, the
player is a heroic character advancing through a fantasy-world setting by
completing a series of multistep tasks, or “quests,” each of which rewards the
player with a certain number of experience points toward the next level of
skill and challenges. The quests themselves are relatively simple—a game
Page 5 of 68
DRAFT; please do not cite or circulate without author’s express permission
character might ask the player to kill ten flying serpents and come back with
their hides, for instance, or to locate a cursed amulet and destroy it in a forge.
But with each new level, the number of points required to “level up” to the
next one rises steeply, so that the highest levels of the game may take many
hours of arduous, repetitive play to traverse.
In the MMO variety of RPGs, the additional presence of thousands
of other players, all potentially interacting, adds further complexities. Tasks
may require banding together with dozens of other players to vanquish
enemies that otherwise cannot be defeated. Acquiring the necessary
resources to advance usually requires trade with other players, sometimes
through barter of one virtual good for another, more often through
payments of the local virtual currency. These interdependencies are typically
reinforced through mandatory specialization of character skills, so that
players particularly adept at mining metals, say, might have to rely on others
better at blacksmithing to turn those metals into armor.
It is important to note, in other words, that MMOs, though games,
are generally not zero-sum games. Except in player-versus-player combat (PvP,
a common feature of MMOs but rarely their main focus), players do not
compete directly with each other. Although each level attained brings new
rewards, in the form of better skills, stronger weapons, and the like, there are
no such rewards for reaching a given level faster than anybody else. Even
informal competition is blunted by the fact that not all players approach the
game as a race to level up. In fact, game designers have long observed that
the highly competitive “achiever” is just one of several player types
Page 6 of 68
DRAFT; please do not cite or circulate without author’s express permission
commonly found in multiplayer RPGs.4 Others are “explorers,” bent on
mapping the game’s virtual world and learning all they can about its
underlying mechanics, or “socializers,” focused on their relationships with
other players, or “killers,” hooked on the thrills of PvP and other forms of
interplayer conflict.5 So disparate are the motivations of these groups that it
opens to question whether they are even playing the same game, let alone
playing it in competition with each other.
Yet even if leveling up is a primary goal only for the achievers, other
player types cannot realistically pursue their own goals without it. Explorers
who don’t level up can never see those parts of the game that are accessible
only to high-level characters; socializers who don’t level up will be left behind
by friends who do; killers who don’t are limited to attacking low-level victims
and are likelier to end up victims themselves. To play an MMO with any
commitment, then, requires a commitment to the path of leveling up. And
given the rigors of the leveling treadmill, that means committing to a baseline
style of play that is at times not obviously playful.
Farming, a typical if not integral aspect of the leveling process, is
MMO play at its most laborious. Broadly, the term—like its approximate
synonym, grinding—refers to any monotonously repeated action undertaken
solely for the purpose of gathering resources useful for advancement in the
game.6 In some MMOs, for example, some portion of the experience points
4
Richard A. Bartle Hearts, Clubs, Diamonds, Spades: Players Who Suit MUDs,
http://www.mud.co.uk/richard/hcds.htm (last visited Feb 16, 2014).
5 Id.
6 BONNIE A. NARDI, MY LIFE AS A NIGHT ELF PRIEST: AN ANTHROPOLOGICAL
ACCOUNT OF WORLD OF WARCRAFT 110 (2010).
Page 7 of 68
DRAFT; please do not cite or circulate without author’s express permission
needed to level up can only be acquired through the routinized slaying of
beasts or monsters (mobs, in MMO parlance) for hours at a time.7 In World
of Warcraft, where it is possible to level up just relying on the somewhat less
onerous mechanism of pre-assigned quests, farming more often means killing
mobs for the gold coins or crafting materials that can be looted from their
corpses.8 Yet there too, because quests and other challenges can often be
completed more efficiently with the help of farmed resources, most players
end up devoting at least some of their leveling time to the tedium of farming.
Perhaps more to the point, however, even players who only do quests
are still engaged in a series of tasks that, as several researchers have observed,
bears a more than passing resemblance to the relatively mindless clickwork of
the modern low-level corporate job.9 It might be easy to conclude that this
resemblance diminishes an MMO’s value as a game—that the more it looks
like work, that is, the less it qualifies as play. Indeed, many players seem to
feel exactly that. A game that compels players to submit to hours of boredom
for the sake of a marginal reward, they complain, is a flawed game.10 Nor are
Grinding, WOWWIKI, http://www.wowwiki.com/Grinding (last visited Feb 16,
2014) (“In older MMORPGs, such as EverQuest, grinding was the primary way to
advance your character's level. The quest system was very underdeveloped, as
opposed to the quest system in World of Warcraft.”).
8 Farming, WOWWIKI, http://www.wowwiki.com/Farming (last visited Feb 16,
2014).
9 See NARDI, supra, note 6 at 110; Scott Rettberg, Corporate Ideology in World of Warcraft
in DIGITAL CULTURE, PLAY, AND IDENTITY: A WORLD OF WARCRAFT READER 19,
eds. Hilde G. Corneliussen and Jill Walker Rettberg (2008); Nick Yee, The Labor of
Fun: How Video Games Blur the Boundaries Between Work and Play, 1 GAMES AND
CULTURE 68 (2006).
10
T.L. TAYLOR, PLAY BETWEEN WORLDS: EXPLORING ONLINE GAME CULTURE
71 (2006). But see also Angry Mustache ALOD: Go back to WoW, THEMITTANI.COM,
http://themittani.com/features/alod-go-back-wow (last visited Apr 15, 2014)
(celebrating as a point of pride the difficulty of the MMO EVE Online as compared
to World of Warcraft).
7
Page 8 of 68
DRAFT; please do not cite or circulate without author’s express permission
they persuaded differently by the fact that a certain class of player, the socalled power gamer, seems to embrace wholeheartedly the worklike aspects
of MMOs, investing upwards of 40-hour weeks in the game and seeking only
to maximize the value of his return in virtual assets. For more casual players,
on the contrary, the power gamer has missed the point of play—to have
fun—and isn’t legitimately playing at all.11
Another view, however, is that the power gamer’s approach is not so
much a rejection of play as a challenge to conventional ideas of what play
looks and feels like. That is the perspective of T.L. Taylor, who has
conducted ethnographic research on power gamers and writes that, in
explaining to her how and why they play, “power gamers do not use the term
‘fun’ . . . but instead talk about the more complicated notions of enjoyment
and reward.”12 Play, for them, is indeed a lot like work, but mainly in the
sense that its pleasures do not vanish on contact with boredom, frustration,
and sober intensity, and may even in some ways depend on them.13
We may take the power gamers at their word on this or not. But if we
accept that play as they experience it verges on indistinguishability from work,
we should keep in mind that this probably distinguishes them only in degree
from most other MMO players. As should be clear by now, the leveling-up
mechanic at the core of MMOs is essentially a simulation of economic
production. And as the next section makes clear, the differences between a
robust economic simulation and an actual economy are fewer and less
TAYLOR, supra note 10 AT 71 (2006).
Id. at 88.
13 Id. at 89.
11
12
Page 9 of 68
DRAFT; please do not cite or circulate without author’s express permission
meaningful than one might think. Given that the very premise of the game
confuses play and productive labor, then, it would be surprising to find that
even its most casual players had not incorporated into their sense of play at
least some trace of the power gamer’s pleasure in the laborious.
B. Farming as Production: The Economies of MMOs
In 2001, economist Edward Castronova calculated the gross
domestic product of what was then the most populous MMO, EverQuest.14
He arrived at the figure through standard econometric methods, inventively
applied. First he looked at dollar-denominated sales of high-level EverQuest
character accounts on real-money auction sites and from those derived the
effective market price of a character’s level (presumed to reflect the value of
both the level attained and the coin, weapons, and armor typically owned by
characters at that level). Then he conducted player surveys to establish the
hourly rate at which the average EverQuest player leveled up. Multiplying
that rate by the price of levels gave him the amount of character wealth an
average player created in an hour. Multiplying that amount by hours in a year,
he got a GDP per capita of $2,226 (a little more than Bulgaria’s). And
multiplying that figure by EverQuest’s player population gave him, finally, a
total GDP of $135 million.
Edward Castronova, Virtual Worlds: A First-Hand Account of Market and Society on the
Cyberian Frontier (CESifo Working Paper Series No. 618, 2001),
http://papers.ssrn.com/abstract=294828 (last visited Mar 24, 2014). Among other
indicators, Castronova also calculated an exchange rate of $0.0107 for EverQuest’s
platinum piece currency (stronger than the yen) and an average hourly wage per
player of $3.42.
14
Page 10 of 68
DRAFT; please do not cite or circulate without author’s express permission
The figure shows statistically what any MMO player knows
intuitively: MMOs may be games, but they are also economies. That they are
virtual economies and not real ones might be a meaningful distinction for
some purposes but not, as Castronova established, for the purposes of basic
economic analysis. Nor is it the real-money market for virtual goods that
makes them valid objects of such analysis. Indeed, even if MMOs had no
internal markets of their own, the leveling treadmill alone would arguably
supply the two basic features of any economic system: scarcity of goods and
a choice of ways to allocate them.15 More than any other feature of the games,
however, it is the simple ability of players to transfer goods among
themselves—found universally in MMOs—that has proved to be the most
economically fateful. Without transferability, there can be no virtual
currencies, virtual markets, real-money trading, or, of course, gold farming.
With transferability, all of these are practically inevitable.
The economic history of MMOs, at least, suggests as much. The
earliest MMOs, dating to the mid-1990s, had transferability and virtual
currencies built in, and organized player-to-player markets arose in them
spontaneously, as buyers and sellers gravitated toward common trading spots
and some began to set themselves up as fulltime in-game merchants.16 Only
later did game designers start routinely building MMOs with markets in mind,
as World of Warcraft’s designers did, for instance, when they located
Roger E. Backhouse and Steven Medema, Retrospectives: On the Definition of
Economics, 23 J. ECON. PERSPECTIVES 221, 225 (2009) (citing a widely accepted
definition of economics as “the science which studies human behavior as a
relationship between ends and scarce means which have alternative uses.”)
16 See Castronova, supra note 14.
15
Page 11 of 68
DRAFT; please do not cite or circulate without author’s express permission
automated auction houses in the capital cities of the game’s fantasy realms.
Real-money trading emerged just as spontaneously—and at least as
predictably, given the number of players short on time or patience for the
grind and happy to buy their way around it. The first trades, presumably,
were informal arrangements among acquaintances, initiated with a check or
cash payment outside the game and sealed with an in-game “gift” of goods
from the seller’s character to the buyer’s. But the rise of auction sites like
eBay and payment systems like PayPal soon ushered in a more commercial
RMT. In-game merchants became real-money merchants, earning
comfortable livings scouring the virtual markets for bargains and selling their
haul at a markup on eBay.17 When the MMO companies started having eBay
shut down virtual-item sales, that only accelerated the commercialization,
pushing RMT retailers onto their own increasingly high-volume ecommerce
sites.
Gold farming was the next step. As RMT retailers consolidated and
their sales volumes grew (fueled in large part by the unprecedented success of
the newly launched World of Warcraft), it grew clear that the cottage industry
of individual players that had thus far been supplying the merchants was too
scattered a production force to keep up. Automation was one response. For a
time, programmers running bot farms—teams of unmanned MMO
characters scripted to harvest virtual resources around the clock—became a
major source of wholesale virtual gold for some retailers.18 But in the end the
See generally JULIAN DIBBELL, PLAY MONEY: OR HOW I QUIT MY DAY JOB AND
MADE MILLIONS SELLING VIRTUAL LOOT (2006).
18 See id.
17
Page 12 of 68
DRAFT; please do not cite or circulate without author’s express permission
economics of offshoring prevailed. The going rate for one hour of a Chinese
worker’s time was roughly $0.25.19 The virtual goods one MMO player could
produce in an hour, meanwhile, were worth easily ten times that amount.20 At
these rates, some bot farmers have complained, a Chinese gold farm’s labor
costs can be competitive even with a U.S. bot farm’s hardware costs.21 Live
labor also makes it possible for gold farms to offer a product bot farms
cannot: “power leveling” services, in which workers temporarily take charge
of customers’ game accounts and level up their characters for them in roundthe-clock bouts of play. These advantages alone perhaps explain why, five
years after the launch of World of Warcraft, a gold farm industry employing
an estimated 400,000 workers had sprung into existence.22
What the logic of labor arbitrage cannot fully explain, however, is
why so much of that industry has been located in China.23 After all, other
countries also have cheap labor and Internet connections (as the gold farms
found from time to time in places like Indonesia, Romania, and Mexico
attest). But when it comes to gold farming, China has at least one advantage
not available to other low-wage countries: Its own vast population of online
gamers, including roughly as many World of Warcraft players as are found in
See Dibbell, supra note 1.
Castronova, supra note 14 (“average wage” of EverQuest players: $3.47).
21 See, e.g., Mithra, Comment to How a Gold Farm Works, TERRA NOVA (Jan. 19, 2006,
5:38 PM)
http://terranova.blogs.com/terra_nova/2006/01/how_a_gold_farm.html#comme
nts.
22 Heeks, supra note 1 at 7.
23 Id. (estimating 85%).
19
20
Page 13 of 68
DRAFT; please do not cite or circulate without author’s express permission
North America and Europe combined.24 This means, for one thing, that gold
farm operators can rely on a work force that needs little on-the-job training.
It also means they can rely on workers who remain, to a surprising extent,
engaged with the job not just as workers but as players. Even after months of
12-hour shifts and 84-hour weeks, for example, gold farmers may still in their
free hours visit the nearest Internet café to play World of Warcraft and other
online games at their leisure. Others, who won’t go near a game after hours
and are frank about their disenchantment with the job, will nonetheless admit
they can’t help feeling the occasional burst of excitement in the midst of the
virtual combat that is their daily routine.25
That even these ostensibly ludic moments redound to the potential
economic benefit of the employer (an excited worker is a focused worker,
after all, and even a worker killing monsters off-the-clock may be refining the
efficiency of his on-the-job technique) might suggest that their playfulness is
only illusory, lacking the defining autonomy of true play. Yet as we saw with
the power gamers, it might also be spurring us toward more complicated
notions of what play can be. Indeed, if we grant that power gaming,
notwithstanding the suspicions of more casual players, is in fact a form of
play, then there remains only narrow ground for insisting that gold farming is
not. Looking purely at their behavioral interactions with the game, for
24
Leigh Alexander, World Of Warcraft Hits 10 Million Subscribers, GAMASUTRA,
http://www.gamasutra.com/php-bin/news_index.php?story=17062 (last visited
Feb 18, 2014) (reporting subscriber figures of 5.5 million in Asia, 2 million in
Europe, 2.5 million in North America).
25 See Julian Dibbell, The Chinese Game Room: Play, Productivity, and Computing at Their
Limits, 2 ARTIFACT 82, 83 (2011)
http://scholarworks.iu.edu/journals/index.php/artifact/article/view/3950/3629.
Page 14 of 68
DRAFT; please do not cite or circulate without author’s express permission
example, one finds little to distinguish the power gamer’s long hours and
maximal efficiency from those of the gold farmer. As for the more subjective
aspects of their gaming: if the power gamer’s stoic endurance of extended
tedium and fatigue for the sake of more remote rewards can count as play,
why can’t the gold farmer’s? Should the fact that the gold farmer’s rewards
are real wages while the power gamer’s are virtual prizes be enough to draw a
firm analytical line between gold farming and the entire universe of play?
Perhaps. Or then again, given how little the distinction between real wealth
and virtual wealth seems to matter for most other purposes, maybe the more
straightforward analysis is to view gold farming, finally, as just another of the
several play styles that commingle in an MMO.
C. Farming as Problem: Gold Farming and the Law
The area of law that speaks to gold farming most directly—and most
frequently—is the law of contracts. Indeed, the job of a gold farmer is, in
one sense, a daily encounter with contract law: Each shift begins with the
gold farmer’s implicitly agreeing, by logging in to a valid MMO account, to
be bound by the provisions of the targeted MMO’s End User License
Agreement (EULA) and Terms of Use (TOU) and proceeds to the extended
violation of those terms that gold farming almost always represents. Realmoney trading and, by extension, gold farming are prohibited by the TOUs
Page 15 of 68
DRAFT; please do not cite or circulate without author’s express permission
or EULAs of all popular MMOs26, and with a decisiveness that contrasts
strikingly with the cultural and economic ambiguities discussed above.
Blizzard’s blunt ban on RMT in World of Warcraft, for example—“[Y]ou
may not sell in-game items or currency for ‘real’ money, or exchange those
items or currency for value outside of the Game”27—does not appear to
leave room for loopholes.
Though the language of these bans is clear, however, the reasons
behind them are less so. The texts of the bans themselves say little or nothing
about what harms they are meant to protect against. Elsewhere, in gamecompany statements and interviews, in user forums and game-news sites, a
handful of explanations has emerged over the years. Game companies
sometimes complain of costs in good will and customer-support time
incurred when RMT deals go bad and players look to the company for
redress. The complaint more commonly heard, especially from players
themselves, is that buying advancement in the game is simply a form of
cheating, or queue jumping. A more sophisticated variation on this argument
targets not so much RMT itself as the gold farming it incentivizes: By
producing more wealth than the game economy was designed to
accommodate, goes the reasoning, the concentrated labor of gold farming
creates in-game inflation that makes it harder for other players to get ahead.28
26
See, e.g., World of Warcraft Terms of Use, BLIZZARD ENTERTAINMENT,
http://us.blizzard.com/en-us/company/legal/wow_tou.html (last visited Apr 15,
2014).
27 Id.
28 Muhammad Aurangzeb Ahmad, Brian Keegan, Jaideep Srivastava, Dmitri
Williams, and Noshir Contractor, Mining for Gold farmers: Automatic Detection of Deviant
Players in MMOGs, in Computational Science and Engineering, 2009. CSE'09. International
Page 16 of 68
DRAFT; please do not cite or circulate without author’s express permission
However, none of these complaints explains, without more, why
game companies should want to ban real-money trading. For one thing, the
bans have costs of their own, including loss of subscription revenues from
gold farms successfully shut down and, perhaps more significant, the
potential loss of players who depend on RMT purchases for their enjoyment
of the game. Even assuming fewer players rely on RMT than revile it as
cheating, that assumption in itself does not make blanket prohibition the
most cost-effective response. In a game without clear winners and losers,
after all, it is not obvious what harms a player causes any other by spending
more money than time to get ahead. It is also, therefore, not obvious that
companies should ban RMT as cheating rather than, as they do with other
controversial play styles, segregate the minority of players who embrace it
from the majority who don’t.29 As for gold farming’s inflationary effects, they
may be real, but they beg the question: Of all the many sources of inflation in
MMO economies—including, notably, the hyperproduction of power
gamers—why prohibit only this one?30 If game companies felt it made
commercial sense to accommodate gold farming, they presumably would do
Conference on, vol. 4, pp. 340-345. IEEE, 2009. (“In-game economies are designed
with activities and products that serve as sinks to remove money from circulation
and prevent inflation. Farmers and gold-buyers inject money into the system
disrupting the economic equilibrium and creating inflationary pressures within the
game economy.”)
29
For the majority of World of Warcraft players who do not enjoy being attacked by
the “killer” type players, for example, unrestricted PvP play is a nuisance more
concrete than RMT; yet rather than ban the killer style altogether, Blizzard has
created separate servers where open PvP is permitted while keeping it restricted on
the rest. At least one company, in fact, has experimented with taking that approach
to RMT. See Michael Zaenke, SOE’s Station Exchange - The Results of a Year of Trading,
GAMASUTRA (February 7, 2007),
http://www.gamasutra.com/view/feature/1716/soes_station_exchange__the_.php
.
30 Castronova, supra note 14.
Page 17 of 68
DRAFT; please do not cite or circulate without author’s express permission
so, factoring it into their fine-tuning of the game’s economy just as they do
with other inflationary phenomena.
Yet even if the business logic of banning RMT remains somewhat
opaque, it seems safe to assume that the logic is there. The almost universal
presence of the ban in MMOs certainly suggests a reasoned conclusion, on
the part of game companies, that the prohibition’s costs and benefits weigh
ultimately in its favor. So too does the companies’ willingness to defend their
RMT bans in court. As early as 2002, in Blacksnow Interactive v. Mythic
Entertainment, Inc.31, MMO developer Mythic Entertainment fended off an
abortive challenge to its RMT prohibition by one of the first known gold
farming operations.32 In Blizzard Entertainment, Inc v. In Game Dollar, LLC33,
Blizzard sued a seller of farmed gold and power-leveling services and, though
the harms alleged were chiefly from the seller’s in-game advertising to World
of Warcraft players, secured an injunction barring the defendant from RMT
of any kind.34 In a more recent farming-related case, the much-cited MDY
Industries, LLC v. Blizzard Entertainment, Inc.35, Blizzard won a judgment
(vacated but not defeated by the Ninth Circuit) against a company that was
Blacksnow Interactive v. Mythic Entertainment, Inc., No. 8:02-cv-00112 (C.D. Ca.
2002).
32
Complaint, Blacksnow at 1 (bringing action for interference with prospective
economic advantage and unfair business practices against MMO developer that had
shut down virtual goods sellers eBay auctions and in-game accounts). See also Julian
Dibbell, Serfing the Web: Black Snow Interactive and the World’s First Virtual Sweat Shop,
WIRED, Jan. 2003, available at http://www.juliandibbell.com/texts/blacksnow.html
(last visited Apr 15, 2014).
33 Blizzard Entertainment, Inc. v. In Game Dollar, LLC, No. 8:07-cv-00589 (C.D.
Cal. 2007).
34 Consent Permanent Injunction at 2, In Game Dollar, No. 8:07-cv-00589.
35 MDY Indus., LLC v. Blizzard Entm't, Inc., 629 F.3d 928 (9th Cir. 2010).
31
Page 18 of 68
DRAFT; please do not cite or circulate without author’s express permission
selling software designed to automate World of Warcraft bots.36 Although
the ban on RMT was not directly at issue in that case, the district court’s
initial ruling noted that Blizzard objected to the software in part because it
facilitated “farming” and real-money trading of virtual assets, “an activity
expressly prohibited by the TOU.”37 That the case turned on an important
question of first impression—namely, the extent to which violation of an
online-service license agreement can trigger provisions of digital copyright—
probably suffices to explain how it reached the appellate level.38 But the fact
that Blizzard fought the case as far as it did hints also at the importance to
game companies of the RMT ban.
Not all attempts at enforcing the ban have originated with the
companies, however. In Antonio Hernandez v. Internet Gaming Entertainment,
Ltd.39, a World of Warcraft player brought a class action suit on behalf of all
non-gold-selling players against what was then the leading retailer of farmed
gold, IGE, naming as co-conspirators all gold farmers who had ever supplied
IGE. Though the parties settled before the court could reach a decision, the
merits of the complaint are worth a closer look. The core harms alleged in it
would have been familiar to anyone who had ever heard World of Warcraft
players complain about the farming and selling of gold: It stoked inflation,
obliging players either to buy gold or to play at a “competitive disadvantage,”
Id. at 957–58. See also Jagex Ltd. v. Impulse Software, 750 F. Supp. 2d 228 (D.
Mass. 2010).
37 MDY Indus., LLC v. Blizzard Entm't, Inc., 2008 WL 2757357 *1 (D. Ariz. July 14,
2008).
38 MDY, 629 F.3d at 954–55.
39 Antonio Hernandez v. Internet Gaming Entm’t, No. 1:07-cv-21403 (S.D. Fla
2008).
36
Page 19 of 68
DRAFT; please do not cite or circulate without author’s express permission
spending far more hours gathering resources for leveling than they would
otherwise have to.40 The central cause of action, meanwhile, was a relatively
simple contractual claim: Namely, that IGE and its co-conspirators breached
their promise under the TOU and EULA not to sell gold and, in so doing,
breached their obligations to all other players as intended third-party
beneficiaries of that promise.41 Yet as Professor Joshua Fairfield has pointed
out, it is not clear that a judge in this case would have found a third-party
beneficiary clause implied in Blizzard’s licensing terms—or, for that matter,
that a bilateral user license agreement can ever really be the right tool for
protecting the complex multilateral interests of an online social world’s
users.42 Here, for example, the Blizzard EULA’s blanket prohibition of realmoney trading doesn’t quite align with Hernandez’s inclusion in his plaintiff
class of players who have bought gold, presumably on the theory that it was
gold farming’s impoverishing macroeconomic effects that obliged them to
“violate the EULA and the ToU by purchasing gold from the
Defendants.”43 Likewise, the inclusion of gold farmers generally as a vast
class of uncharged co-conspirators further strains the limits of the
contractual framework. It’s almost as if what Hernandez was seeking wasn’t
so much a private law remedy as something more broadly social—a judicial
refereeing of the conflict between two competing modes of organizing the
Amended Class Action Complaint, Hernandez, No. 1:07-cv-21403.
Id.
42 Joshua A.T. Fairfield, Anti-Social Contracts: The Contractual Governance of Virtual
Worlds, 53 MCGILL L.J. 427, 453-54 (2008).
43 Id.
40
41
Page 20 of 68
DRAFT; please do not cite or circulate without author’s express permission
productive activity of players, perhaps: A sort of labor and employment law
of the MMO.
The legal scholarship relevant to RMT and gold farming traces
roughly the same conceptual contours as the case law. There is just a lot
more of it. Legal-academic fascination with virtual worlds in general and
virtual economies in particular has generated hundreds of articles in the last
decade.44 Among those, the topic probably addressed most often is that of
virtual property rights.45 This focus in part reflects attention drawn to the
open virtual world Second Life, where RMT has never been prohibited and
individual users have been known to amass holdings of virtual assets valued
in the millions of dollars. But the question whether MMO players retain any
rights to in their in-game assets other than what the game company grants
them by agreement—the question looming in every legal challenge to the ban
on RMT, that is—has also accounted much of the scholarship on the subject.
Look for articles discussing virtual work, on the other hand, and you will find
a dozen at most, and then only if you count the handful of articles that have
taken up the question whether virtual items acquired through in-game efforts
should be taxed as income.46 Leave those out, and the number of articles
viewing virtual worlds through the lens of labor appears to be exactly two,
both by Professor Miriam Cherry, and both intent on how U.S. employment
A WestLaw search on “virtual +1 worlds” as of March 12, 2014, returns 1062
articles and treatise entries.
45 A WestLaw search on “virtual +1 property” as of March 12, 2014, returns 434
articles and treatise entries.
46 See, e.g., Theodore P. Seto, When Is a Game Only a Game?: The Taxation of Virtual
Worlds, 77 U. Cin. L. Rev. 1027, 1029-30 (2009); Bryan T. Camp, The Play’s the Thing:
A Theory of Taxing Virtual Worlds, 59 Hastings L.J. 1 (2007); Leandra Lederman,
“Stranger Than Fiction”: Taxing Virtual Worlds, 82 N.Y.U. L. Rev. 1620 (2007).
44
Page 21 of 68
DRAFT; please do not cite or circulate without author’s express permission
and labor laws might apply to a range of activities in and around virtual
worlds.47
This article in a sense just picks up where Cherry’s comments leave
off. Unlike her work, however—and that of most other commentators on
virtual worlds—this article looks only at MMOs, to the exclusion of Second
Life and other virtual worlds marketed more as open-ended platforms than
as games. The point is not so much to narrow the scope of the analysis as to
sharpen its focus. The point, that is, is to keep the element of play front and
center and, in so doing, connect the literature on virtual economies to work
on the commodification of play and, through it, to debates on
commodification more broadly.
II. COMMODIFICATION AND PLAY
A. The Commodification Debate
Norms against taking or giving money in exchange for the
performance of an otherwise licit activity are, of course, a phenomenon
much older than the bans on gold farming. Ancient examples include the
Bible’s strictures against prostitution and usury.48 Newer restrictions have
emerged as markets, technologies, and cultural practices have evolved. Babies
and human body parts, for example, now number in the class of things that,
Miriam A. Cherry, Working for (Virtually) Minimum Wage: Applying the Fair Labor
Standards Act in Cyberspace, 60 ALA. L. REV. 1077 (2009); Miriam A. Cherry, A
Taxonomy of Virtual Work, 45 GA. L. REV. 951, 993-94 (2011).
48 See Deuteronomy 23:17–18 (prostitution); Ezekiel 18:7–9 (usury).
47
Page 22 of 68
DRAFT; please do not cite or circulate without author’s express permission
for the most part, may be given but not sold. Likewise, the persistence of
term limits on copyright, the denial of patents on human genes, and other
checks on the modern expansion of intellectual property rights can be taken
to express a judgment that some kinds of information are (or should be)
priceless.
The target of all these prohibitions is what in some academic circles is
called commodification. Among legal scholars, commodification has been a
focus of debate for many years.49 Pro-market law-and-economics thinkers
gave an early impetus to the discussion. Judge Richard Posner’s argument in
favor of an open market in babies, for example, was a particularly
provocative conversation starter, challenging defenders of adoption law’s
status quo to show that payments from adoptive to birth parents would not
in fact produce better outcomes for all involved, including the adopted
children.50 The debate ever since has largely been, on some level, a response
to the Chicago school and its commitment to what Professor Margaret Jane
Radin has called “universal commodification” as both analytic tool and policy
preference.51
Radin’s own work on the subject is an extended, nuanced critique of
universalism. For Radin, there are two main concerns that arguments like
Posner’s fail to fully account for. One, resting on what Radin calls a “domino
See generally RETHINKING COMMODIFICATION: CASES AND READINGS IN LAW
(Martha M. Ertman and Joan C. Williams, eds., 2005).
50 Elisabeth M. Landes & Richard A. Posner, The Economics of the Baby Shortage, The, 7
J. LEGAL STUD. 323 (1978).
51 MARGARET JANE RADIN, CONTESTED COMMODITIES: THE TROUBLE WITH
TRADE IN SEX, CHILDREN, BODY PARTS AND OTHER THINGS 2 (1996) (“Our
investigation of contested commodification must begin with an understanding of
the archetype in which commodification is uncontested.”).
49
AND CULTURE
Page 23 of 68
DRAFT; please do not cite or circulate without author’s express permission
theory” of commodification, is the concern that creating markets in even a
limited subclass of intimately personal interactions will cause market logics to
spread corrosively to comparable but as-yet uncommodified interactions.52
For example, even if on balance adopted children and their parents are better
off with robust baby markets in place, their gains may not fully compensate
for the harms, dignitary and otherwise, that could befall all children once they
and their families know to the dollar how the market values any child’s
particular attributes.53 The other concern is that, even if the mere act of
pricing the priceless in itself does only the airiest of harms, in the world as it
exists those harms inevitably are interwoven with—and magnified by—the
effects of poverty, racism, sexism, and other “wrongful subordinations” on
would-be sellers of their own commodified personhood.54
Radin’s arguments have in turn been challenged by another set of
commentators. Drawing on the insights and observations of economic
sociology, these scholars question the notion, implied in Radin’s domino
theory, that there are realms of the social that are or even could be kept
wholly free of commodification. “Actual studies of concrete social settings,
from auctions to households, do not yield descriptions of spheres neatly
separated,” write legal scholar Patricia Williams and sociologist Viviana
Zelizer.55 Instead, they explain, we see “dense networks of social relations
that intertwine the intimate and economic dimensions of life”: markets and
Id. at 95.
Id. at 100.
54 Id. at 163–2.
55 Joan C. Williams and Viviana A. Zelizer. To Commodify or Not to Commodify: That Is
Not the Question, in RETHINKING COMMODIFICATION: CASES AND READINGS IN
LAW AND CULTURE at 367 (2005).
52
53
Page 24 of 68
DRAFT; please do not cite or circulate without author’s express permission
corporations sustained by personal ties, family homes shot through with
economic relations both internal and external.56 Noah Zatz, in his work on
prison labor and employment law, extends the argument to show how even
seemingly crisp distinctions between market and nonmarket activities can
blur under close inspection—as when the freedom of participation that
might be presumed to distinguish market labor from prison labor reveals
itself to be less than total, compromised by the inequality of workplace
power relations, the coercive threat of unemployment, and other elements of
the “backdrop of compulsion” against which ordinary employees work.57
For Zatz, the instability of the distinction between markets and
nonmarkets has a clear methodological implication: It suggests abandoning
the distinction as a tool of analysis and making it instead an object of study,
examining the lawmaking and other social processes that, under the guise of
policing the line between markets and other social spheres, are often in fact
constructing it.58 For Williams and Zelizer, similarly, the belief in a clear
separation of markets and nonmarkets rests not on empirical reality but on
liberal and patriarchal ideologies whose pedigree dates to the origins of
industrial capitalism. There is a long and impressive intellectual history, they
note, of dividing social reality into discrete, mutually antagonistic realms of
possessive individualism, on the one hand, and domestic communalism on
the other—of rationality and sentiment, of self-interest and solidarity, of
Id. at 366.
Noah D. Zatz, Prison Labor and the Paradox of Paid Nonmarket Work, in ECONOMIC
SOCIOLOGY OF WORK at 25 (Nina Bandelj ed., 2009). [Vol. 18 of Research in the
Sociology of Work.] [369–98]
58 Id. at 4.
56
57
Page 25 of 68
DRAFT; please do not cite or circulate without author’s express permission
Gesellschaft and Gemeinschaft. This way of looking at the world has shaped the
world we know, Williams and Zelizer acknowledge, but it is not the only
plausible way, nor is it, they argue, the most accurate.
In rejecting this history of dualisms, commodification theory’s
sociological turn seeks also to reframe the commodification debate as more
than a binary choice between resisting commodification and embracing it.
Yet ultimately there is a conflict of interests at the core of the debate that is
not so easily rationalized away. Consider the “double bind” that Radin poses
as a paradigmatic hard problem for commodification theory: That of the
impoverished person whom hunger has pushed to the brink of prostituting
herself or selling a kidney. If the law stops her, she may starve. If the law
permits her to proceed, she risks harms to her health and dignity. In neither
case does she fully have a choice; in neither case can we be happy with the
consequences of the law. The dilemma is gripping, but for Williams and
Zelizer, and as Radin herself acknowledges, it may also be false, an artifact of
the compulsion to judge commodification rather than the broader conditions
of oppression that give rise to hunger and marginalization.59
Yet Radin is right, rhetorically at least, to focus on the double bind,
because it throws into relief a tension that has fueled the debate so far and
that cannot be written off as a figment of the liberal imaginary. The tension is
this: On the one hand, individuals have an interest in the right to maximize
the value of a personal attribute, whether it is their intelligence or a bodily
organ, and whether they realize its highest value by using it themselves or by
59
Williams, supra note 55, at 373; RADIN, supra note 51 at 123–30.
Page 26 of 68
DRAFT; please do not cite or circulate without author’s express permission
trading it away. On the other hand, they have an equally powerful interest in
the integrity of their selves and of the ties that bind them, out of more than
the immediate self-interest of commercial exchange, to other selves. The one
interest thrives in market settings and the other does not, yet neither is purely
economic or purely personal. There is a dignity in the autonomy that
freedom to participate in markets to the fullest represents, just as there are
material benefits to social solidarity and psychic well-being. They are the
defining poles of the commodification debate, and as we will see, they are
just as critical to understanding what’s at stake when the human attribute
commodified is the capacity for play.
B. The Commodification of Play
1. Play theories and commodification
In a list of things closely related to human beings’ sense of
personhood and community (and therefore likely to be governed by
regulatory regimes that limit their circulation as market goods), Radin
includes the following: “homes, work, food, environment, education,
communication, health, bodily integrity, sexuality, family life, and political
life.”60 Why play is not on the list Radin does not say, but a good guess would
be that it simply did not cross her mind. This would certainly be consistent
60
RADIN, supra note at 113.
Page 27 of 68
DRAFT; please do not cite or circulate without author’s express permission
with the fact that, as a topic of discussion, play comes up nowhere in the
remainder of Radin’s works on commodification. And that fact, in turn, is
entirely consistent with play’s more or less complete absence from the body
of legal scholarship with which those works are in conversation.
The absence is puzzling. In other scholarly realms the
commodification of play has been a topic of at least oblique concern for
nearly a century, and not surprisingly. In the conventional modern view, after
all, play is not just an essentially noneconomic activity but, arguably, the
essentially noneconomic activity. That we conventionally speak of work and
play as opposites is perhaps the bluntest cultural expression of this notion.
But sophisticated theories of play have also tended to support it. Johan
Huizinga, in his classic cultural history of play Homo Ludens61, declared it a
defining characteristic of play that it is “connected with no material interest,
and no profit can be gained by it.”62 Taking the definition further, the
sociologist Roger Caillois called play “an occasion of pure waste”63 and
warned of what follows when it becomes productive of real-world
consequences:
What was an escape becomes an obligation, and what was a pastime
is now a passion, compulsion, and source of anxiety. The principle of
play has become corrupted. It is now necessary to take precautions
JOHAN HUIZINGA, HOMO LUDENS: A STUDY OF THE PLAY-ELEMENT IN
CULTURE. (1955).
62 Id. at 13.
63
ROGER CAILLOIS, MAN, PLAY, AND GAMES 5 (Meyer Barash, trans., 2001).
61
Page 28 of 68
DRAFT; please do not cite or circulate without author’s express permission
against cheats and professional players, a unique product of the
contagion of reality.64
Along similar lines, the philosopher Bernard Suits emphasized the essential
inefficiency of play and games, citing the example of foot race contestants
circling around a track rather than cutting more directly across the field to
the finish line.65 For Suits, this thoroughgoing commitment to the
nonproductiveness of play was the key ingredient in what he called the
“lusory attitude” that distinguishes playful from other kinds of exertions.
More recent thinking about play, however, has begun to question the
insistence on separating it from productivity. In part, and particularly through
the work of play theorist Brian Sutton-Smith, this reconsideration of play’s
productive dimensions has resulted from careful review of developments in
the philosophy, psychology, and ethnography of games and play.66 Among
those developments, special mention perhaps should go to psychologist
Mihaly Csikszentmihalyi’s research on “flow,” an exhilarating state of
hyperfocused activity that appears to happen to people as often and as
intensely on the job as in the midst of play.67 But as much if not more of the
impetus to rethink the work/play opposition has come not from combing
through the literature but from researchers’ direct experience of a burgeoning
Id. at 44–45.
BERNARD SUITS, THE GRASSHOPPER: GAMES, LIFE AND UTOPIA (2005).
66
BRIAN SUTTON-SMITH, THE AMBIGUITY OF PLAY 173–213 (1997). See also
Thomas M. Malaby, Anthropology and Play: The Contours of Playful Experience, 40 NEW
LITERARY HISTORY 205 (2009).
67
MIHALY CSIKSZENTMIHALYI, FLOW: THE PSYCHOLOGY OF OPTIMAL
EXPERIENCE 143–63 (2008) (chapter on “Work as Flow”).
64
65
Page 29 of 68
DRAFT; please do not cite or circulate without author’s express permission
landscape of hardcore video games that ranges from deceptively simple
exercises in mobile-app masochism like Flappy Bird to the level grinding,
power gaming, and real-money trading of MMOs.68 The more play theorists
see of what play can be, the firmer they appear to grow in their conviction
that the once-sacrosanct conceptual firewall between play and productivity
has really, all along, been just a relic of capitalist ideological history: a fourhundred-year-old “Puritan play ethic” (as Sutton-Smith calls it) designed to
elevate the post-Calvinist virtues of accumulative labor by slandering play as
utter uselessness.69
Having moved swiftly from its initial faith in industrial-age dualisms
to a more nuanced sense of the intertwining of economic and noneconomic
dimensions within a single social realm, play theory has traced an intellectual
trajectory almost identical to the one described above for commodification
theory—and the similarity can hardly be accidental. In fact it’s not: The
arguments play theorists have been working through are more than just
analogous to the commodification debate, they are part of it.
This can be difficult to recognize, when so much of the
commodification that stirs gamers to debate falls outside familiar patterns.
Real-money trading is of course a classic encroachment of markets on
nonmarket space. But online games provide many opportunities for markets
to make themselves felt without money ever changing hands. As we saw
when discussing the phenomenon of MMO power gamers, for example, the
68
See Celia Pearce, Productive Play: Game Culture From the Bottom Up, 1 GAMES &
CULTURE 17 (2006); Yee, supra, note 6.
69 SUTTON-SMITH, supra note 66, at 201.
Page 30 of 68
DRAFT; please do not cite or circulate without author’s express permission
power gamers’ playing hours and task pacing can reach levels that, especially
in combination with the economic mechanisms that permeate the game, can
feel like those of a paid job and provoke some of the same anxiety,
compulsion, and sense of obligation Caillois identified as symptoms of play’s
contamination by the “contagion of reality.” At what point are there enough
indicia of commodification that we can declare, even absent any real-money
transaction, that commodification has occurred? (And would we do so even
knowing that a power gamer may profess to take a certain pleasure in the
stress and intensity of his playing style?)
Spillover effects are another way for market dynamics to make
themselves felt. Assuming we count the power gamer’s play style as an
instance of commodification in itself, then the more casual player who feels
pressured to keep up with that pace has also been touched by
commodification. Such effects can be more acute in other types of games.
Social-media-based leveling games like the former Facebook hit FarmVille,
for instance, are typically played for free by the majority of players, with
revenues derived both from advertising and from in-game sales of power-ups
to players impatient to reach the next level. In principle, those who play for
free need never think of themselves as customers, but in practice they may
feel the effects of the business model one way or another. Incentivized to sell
more power-ups, the game company will tend to slow the leveling pace or
otherwise jigger the game mechanics in attempts to wear down the patience
of the highest-level free players. The ads channel, likewise, will incentivize
the company to keep players in the game for as long as possible, leading to
Page 31 of 68
DRAFT; please do not cite or circulate without author’s express permission
more addictive game mechanics than the player of a subscription-fee-based
game would encounter.
Other revenue models pose yet more challenging commodification
scenarios. In some games, for example, revenues come neither from ads nor
from sales to players but from third party purchases of player services. That is,
players who want a power-up will first have to, say, fill out a customer’s
survey or give feedback on a customer’s product.70 In these games the player
steps outside the game to perform the revenue-generating actions required to
continue with the game. In some variations, though, known by the generic
term gamification, playing the game is the required action.71 In one such
experiment, Google offered a moderately addictive online team-based,
matching game in which teams were presented with images from Google’s
database and earned points if, without colluding, they came up with the same
terms as labels for the images. Players knew that in playing the game they
were performing a service for Google: Coming up with functioning image
tags at a rate that artificial intelligence was still not capable of matching.72
Players didn’t have to know this to enjoy the game, however. Could we even
call it commodified play if only Google felt the economic effects of the
transaction? What about other instances of gamification? The web site that
'Crowdsourcing' Employer Denies Minimum-Wage Violations: Otey v. CrowdFlower Inc.,
2013 WL 444500 *1 (WESTLAW J. EMP.) (quoting CrowdFlower executive: “I love
it because we almost trick the game players into doing something useful for the
world while playing these games. Just to do 10 minutes of real work that a real
company can use and we'll give you a virtual tractor”).
71 See Miriam A. Cherry, The Gamification of Work, 40 HOFSTRA L. REV. 851 (2011).
72 Google Image Labeler, WIKIPEDIA, THE FREE ENCYCLOPEDIA (2014),
http://en.wikipedia.org/w/index.php?title=Google_Image_Labeler&oldid=594327
595 (last visited Apr 15, 2014).
70
Page 32 of 68
DRAFT; please do not cite or circulate without author’s express permission
turns children’s household chores and homework into a leveling game?73
Does that cross the line into commodification of play, even if there are no
revenues?
All of this is to say that play poses a rich set of questions for
commodification theory, which would thus be well served by the
incorporation of play into its list of core problem areas. As for play theorists,
now that they have rejected the Puritan model of play as pure waste, they are
faced with a model that provides more flexibility but also more complexity.
Before, any activity that had an economic dimension was not play. But what
is play now? In their own attempts to move beyond simple binarisms, play
theorists, in their turn, would do well to recognize their common intellectual
cause with the theorists of commodification. And to begin with, play theory
might do best to look more closely than it has at commodification theory’s
primary source for social understandings of the relationship between markets
and fundamental human attributes: The law.74
2. Commodified play and the law
CHORE WARS, http://www.chorewars.com/help.php (last visited Apr 15, 2014)
(“Chore Wars lets you claim experience points for household chores. By getting
other people in your house or workplace to sign up to the site, you can assign
experience point rewards to individual tasks and chores, and see how quickly each
of you levels up.).
74 A notable exception to play theory’s general lack of attention to case law is Greg
Lastowka’s chapter on “Games” in his Virtual Justice: The New Laws of Online Worlds,
to which my own discussion in the following section is greatly indebted. See GREG
LASTOWKA, VIRTUAL JUSTICE: THE NEW LAWS OF ONLINE WORLDS 103–21
(2010). Huizinga also famously discusses law in Homo Ludens, but he does so only to
compare its systemic similarities to—and ancient roots in—play. HUIZINGA, supra
note 61.
73
Page 33 of 68
DRAFT; please do not cite or circulate without author’s express permission
Unlike both classic play theory and conventional wisdom, the law
seems to recognize no stark distinction between work and play. Certainly it
has no trouble acknowledging that a player can be a worker—as indicated,
for example, by the well-established recognition of professional athletes’
unions under labor laws.75 Likewise, as rulings granting employee status to
student athletes have shown, even a rigorously policed amateurism may not
suffice to keep courts from assimilating play to the laws that govern work.76
Conversely, nothing in these precedents or any others rules out the
proposition that what is work can still be play. Indeed, whatever Caillois’s
doubts about the corrupting influence of pay on play, strong norms against
the “throwing” of professional games evince a general expectation that pro
athletes are not merely mimicking play when they do their jobs—and the
law’s support for such norms, as for example in the legal aftermath of the
1919 “Black Sox” scandal, implicitly endorses that expectation.77 By the same
See, e.g., Michael H. LeRoy, Federal Jurisdiction in Sports Labor Disputes, 2012 UTAH L.
REV. 815 (2012) (“Textual and data analysis of eighty-three published opinions from
1970-2011 shows that professional athletes . . . used [ ] federal laws to achieve
greater labor market mobility. Under the National Labor Relations Act (NLRA),
players formed unions and bargained collectively with leagues.”).
76 See, e.g., Robert A. McCormick, Amy Christian McCormick, The Myth of the StudentAthlete: The College Athlete As Employee, 81 WASH. L. REV. 71 (2006). See also Alejandra
Cancino, Northwestern, Football Players, Cap Off NLRB Arguments in Union Effort,
CHICAGO TRIBUNE (Mar. 18, 2014), http://articles.chicagotribune.com/2014-0318/business/chi-northwestern-football-union-closing-brief-20140318_1_footballplayers-former-players-graduate-head-football-coach (discussing NLRB’s hearing of
a petition by Northwestern University football players for recognition of their right
to unionize). Michael McCann, Two Potentially Game-Changing Developments in O’Bannon
vs. NCAA, SI.COM (DEC. 17, 2013), http://sportsillustrated.cnn.com/collegefootball/news/20131217/obannon-case-ncaa-latest-developments/ (last visited Mar.
24, 2014) (discussing N.D. Cal. ruling allowing former student athletes to proceed
with suit demanding shares of profits from broadcasts and licensing of their play
and likenesses).
77 See Douglas Linder, The Black Sox Trial: An Account,
http://law2.umkc.edu/faculty/projects/ftrials/blacksox/blacksoxaccount.html (last
75
Page 34 of 68
DRAFT; please do not cite or circulate without author’s express permission
token, if the law really accepted the mutual exclusivity of work and play, it
would have to exclude from its definitions of employment not only any
professional player who takes her job seriously but, arguably, anyone who has
fun doing her job.
This is not to say that the law is incapable of drawing lines between
activity that is play and activity that is not. When called upon to do so,
however, courts tend to focus not on the presence or absence of a “lusory
attitude” in the participants (a difficult inquiry even granting the
distinguishability of such an attitude from the “flow” state of a happy
worker) but on the arbitrary rules that are an equally pervasive feature of play.
When intervening in disputes arising from the midst of organized game play,
such as negligence suits over unnecessary roughness or bad referee calls on
the football field, courts have often deferred to the game as if to a parallel
jurisdiction, presuming to pass judgment only on behavior that did not
constitute play as the game’s rules defined it.78 Where courts have presumed
to pass judgment on a rule itself, they have done so on the similarly
deferential premise that they judged only whether the rule was actually a rule
of the game in question.
visited Mar 24, 2014) (detailing the trial and acquittal of eight Chicago White Sox
players on charges of fraud for throwing the 1919 World Series). But see also
Matthew B. Pachman, Note, Limits on the Discretionary Powers of Professional Sports
Commissioners: A Historical and Legal Analysis of Issues Raised by the Pete Rose Controversy,
76 VA. L. REV. 1409, 1414 (1990) (discussing legal limits on the power of governing
athletic bodies to enforce rules including those against gambling on games by
participants).
78 See Hackbart v. Cincinnati Bengals, Inc., 601 F.2d 516, 520 (10th Cir. 1979)
(finding a football player’s crippling backhand blow to another’s head mid-game to
be actionable only because it exceeded the degree of violence permitted by football’s
rules). See also LASTOWKA, supra note 74, at 105–13 (citing Hackbart and other cases
to argue that “the gulf between law and games is . . . due to the fact that games
constitute a rival regime of social ordering”).
Page 35 of 68
DRAFT; please do not cite or circulate without author’s express permission
In deciding which rules are and are not game rules, courts have
looked to a variety of factors. In some cases, the question has been who
imposed the rule. In Uston v. Resorts Int’l Hotel, Inc. 79, the New Jersey Supreme
Court held that the state’s Casino Control Commission had sole authority to
make the rules of casino games—and that an Atlantic City casino’s rule
against card counting in blackjack, unsupported by any corresponding rule
set by the Commission, was therefore unenforceable against a professional
card counter.80 In other cases, courts have inquired more directly into the
relationship between the rule and the game. In PGA Tour Inc. v. Martin81, for
example, the U.S. Supreme Court invalidated the Professional Golf
Association’s rule requiring golfers to walk between holes rather than take a
cart on the grounds that it was not “an essential attribute of the game
itself.”82 The reasoning in Martin drew in part on historical accounts of how
golf has always been played.83 But in videogame clone cases like Tetris Holding,
LLC v. Xio Interactive, Inc.84, courts have had to make more purely formal
decisions about what aspects of a computer game are embodiments of rules
(and therefore not copyrightable) and which are aesthetic features.85
Uston v. Resorts Int’l Hotel, Inc., 89 N.J. 163 (1982).
Id. at 166. By counting the number and face values of cards dealt by a casino
blackjack dealer, players can gain an advantage over the house that effectively
assures them a profit from long-term play. In this and other aspects of their
livelihood, professional card counters are perhaps the nearest analogue to gold
farmers outside of the MMO context.
81 PGA Tour Inc. v. Martin, 532 U.S. 661 (2001).
82 Id. at 685.
83 Id. at 683–85.
84 Tetris Holding, LLC v. Xio Interactive, Inc., 863 F. Supp. 2d 394 (D.N.J. 2012).
85 See id. at 404 (“The game mechanics and the rules are not entitled to protection,
but courts have found expressive elements copyrightable, including game labels,
design of game boards, playing cards and graphical works.”). See also Bruce E.
79
80
Page 36 of 68
DRAFT; please do not cite or circulate without author’s express permission
If the law seems willing to accept in principle, therefore, that play is
at times indistinguishable from work, it seems equally amenable to the task of
protecting play from the “contagion of reality” if need be. In the context of
online games and their economies, this structural ambiguity has led some
commentators to call, perhaps impatiently, for clarification of the law’s
position on the commodification of online play. On the one hand, Joshua
Fairfield has advocated for recognition of robust player property rights in
virtual assets held not only in Second Life–type worlds but in MMOs,
arguing that the standard end-user contract is an insufficient mechanism for
protecting players’ interests in the fruits of their play.86 On the other hand,
Edward Castronova, disturbed by the implications of his own economic
analyses, and echoing Margaret Radin’s concerns about runaway
commodification, has argued that the transmutability of virtual commodities
into real ones threatens the function of virtual worlds as havens of play and
fantasy.87 Castronova shares with Fairfield a skepticism that the EULA is
adequate to protect players’ interests, in this case their interests in guarding
against the domino effects of commodification.88 He proposes strengthening
the EULA through creation of a doctrine of “interration” for virtual worlds,
analogous to that of incorporation for organizations, which would limit the
Boyden, Games and Other Uncopyrightable Systems, 18 GEO. MASON L. REV. 439 (2011)
(elucidating the rationale for the exclusion of game rules from copyright).
86 Joshua A.T. Fairfield, Virtual Property, 85 B.U. L. REV. 1047 (2005). But see, among
others, Steven J. Horowitz, Competing Lockean Claims to Virtual Property, 20 HARV. J.L.
& TECH. 443 (2007) (arguing that any property claims players might advance in a
game’s assets must remain subordinate to those of the game developer).
87 Edward Castronova, The Right to Play, 49 N. Y. L. SCH. L. REV. 185, 196 (2004).
88 Id. at 197.
Page 37 of 68
DRAFT; please do not cite or circulate without author’s express permission
reach of property, tax, employment, and other primarily economic laws into
these play spaces.89
To propose, instead, shifting the analytical focus away from questions
of property and contract and onto the fault line between work and play is not
to presume that doing so bypasses the disagreement between Fairfield and
Castronova. Their respective concerns effectively recapitulate the
commodification debate’s fundamental tension—between validating the
interest in the market alienability of intimate attributes, on the one hand, and
preserving the integrity of personal and communal space on the other—and
a mere reframing isn’t likely to resolve it. Moreover, as should be clear by
now, the line between work and play is hardly, in itself, a fount of clarity.
Nonetheless, by narrowing the question to what is and is not different
between gold farmers and their fellow online gamers, we gain a perspective
on the problem of commodified play that may afford both new insights and
new solutions.
III. ARE UNPAID FARMERS EMPLOYEES?
This article presumes that gold farming meets the definition of
employment necessary for coverage under U.S. labor and employment laws.
The first question it asks in distinguishing gold farmers from their unpaid
Id. at 201. Other discussions of virtual worlds have similarly argued for protecting
them from the impingements of economic law. See, e.g. Camp, supra note 46, at 69
(arguing for qualified exemption of virtual economies from income taxation as a
response to “the feared commodification of virtual worlds”).
89
Page 38 of 68
DRAFT; please do not cite or circulate without author’s express permission
counterparts, therefore, is the extent to which that definition also covers
unpaid farming. As mentioned above, this inquiry relies in no small measure
on Miriam Cherry’s analysis of employment law as it applies to virtual work.90
In some ways, however, Cherry’s treatment of the question is at once too
broad and too narrow to serve this article’s purposes. It is broad in that
Cherry’s focus extends beyond the uniquely ludic space of MMOs, to Second
Life and other virtual workplaces in which the element of play has a less than
central role.91 It is narrow, on the other hand, in not extending to varieties of
work that might, by way of comparison, situate gold farming within the
commodification debate more generally.
Apposite points of comparison are not difficult to identify. The
contested employment status of student athletes is, as we have seen, an
obviously relevant case of ambiguously commodified play. Sex work, too,
comes readily to mind as the commodification of a pleasure that is as
peculiarly personal, in many ways, as that of play. Yet the singularly economic
character of MMO play complicates these comparisons. Leave the money out
of sport, sex, domestic care, and other canonically contested commodities,
and what remains are not obviously economic activities. Keep the RMT out
of an MMO, however, and it remains essentially a game of productive
accumulation and market exchange. In this, MMO farming invites
comparison not so much to other forms of commodified or commodifiable
See, in particular, Cherry (2009), supra note 47.
See id. at 1089 (discussing the “crowdsourcing” site Amazon Mechanical Turk, a
virtual labor market in which data entry, metatagging, and other microtasks are
assigned and paid for). See also Amazon Mechanical Turk,
https://www.mturk.com/mturk/welcome (last visited Mar 25, 2014).
90
91
Page 39 of 68
DRAFT; please do not cite or circulate without author’s express permission
play as, somewhat surprisingly, to prison or slave labor, in which the laborer
performs conventionally economic activities (construction, manufacture,
farm work) but for unsettlingly unconventional motives.92
For this and other reasons, therefore, the analysis in this section
rounds out its frame of reference by looking to modern case law on prison
labor, and in particular to Noah Zatz’s insightful discussions of the law’s
struggles to resolve the employment status of prison laborers.93
A. Do Players Meet the Control Test of Employment?
The FLSA defines an “employee” as “any individual employed by an
employer.”94 To “employ,” in turn, is to “suffer or permit to work.”95 As
federal statutory definitions of employment go, this one is fairly typical in its
opacity.96 Faced with such definitions, courts fill in the blanks by reading
them to incorporate the common-law test for employment. This test derives
from traditional agency law and focuses on the “the hiring party’s right to
The rumor that Chinese prisons have long enlisted inmates in gold farming
operations lends no particular force to this comparison but is hard to resist passing
along. See Danny Vincent, China Used Prisoners in Lucrative Internet Gaming Work, THE
GUARDIAN (May 25, 2011),
http://www.theguardian.com/world/2011/may/25/china-prisoners-internetgaming-scam (quoting one ex-prisoner’s recollection of forced gold farming: “If I
couldn’t complete my work quota, they would punish me physically. . . . We kept
playing until we could barely see things”).
93 See Noah D. Zatz, Working at the Boundaries of Markets: Prison Labor and the Economic
Dimension of Employment Relationships, 61 VAND. L. REV. 857 (2008). See also Zatz,
supra note 57.
94 29 U.S.C. § 203(e)(1) (2006).
95 29 U.S.C. § 203(g) (2006).
96 See Zatz, supra note 93, at 871 n.50 (citing the Equal Employment Opportunity
Act’s similarly “brief, vague” definition of employment).
92
Page 40 of 68
DRAFT; please do not cite or circulate without author’s express permission
control” the worker and her work.97 Courts have developed a long,
nonexhaustive list of factors for determining whether a worker is under such
control and therefore an employee, rather than an independent contractor.98
A finding of employee status grows likelier, for example, when the employer
has the power to direct the way work is done, to set the hours worked, or to
hire and fire, when the employer provides work benefits, when the employer
(rather than the worker) supplies the requisite equipment, when the work
relationship between the parties is permanent or relatively long-lasting, when
the work requires low skills and little training, when pay is tied to time
worked rather than projects completed.99
In applying the control test to virtual work, Cherry appears not to
contemplate the possibility that the test could identify a relationship of
employment between an MMO’s developer and its players.100 This may well
be a simple oversight, since, as we will see, Cherry’s subsequent discussion of
MMO item farming seems to presume that if farming is work covered by the
FLSA, then it is the MMO’s developer that the farmer effectively works for.101
Yet if Cherry means to imply, instead, that courts are likely to be wary of
Nationwide Mut. Ins. Co. v. Darden, 503 U.S. 318, 322-23 (1992).
See id. at 323–24 (providing a nonexclusive list of 12 such factors).
99 See Herman v. Express Sixty-Minutes Delivery Serv., Inc., 161 F.3d 299, 303–305
(5th Cir. 1998); Martinez-Mendoza v. Champion Int’l Corp., 340 F.3d 1200, 1209
(11th Cir. 2003).
100 Cherry’s control-test discussion focuses, instead, on the documented relationship
between a marketing company operating inside Second Life and the Second Life
users it hired to work shifts as “greeters” in its virtual retail shops. Noting that the
Internal Revenue Service already issued a private letter ruling that, under a controltest analysis, the greeters were employees for the purposes of Social Security tax
laws, Cherry nonetheless concludes that their employee status under the FLSA
would be “a close question.” Cherry, supra note 47, at 1096–98.
101 Id. at 1102 (“[W]hile performing a task may not directly benefit the company, it
might provide an indirect benefit. . . . [that] could attract more users to the world”).
97
98
Page 41 of 68
DRAFT; please do not cite or circulate without author’s express permission
claims that game companies control players to the same extent employers
control employees, her doubt is not unreasonable. Considering some of the
work regimes that have been found insufficiently controlling to amount to
employment—like the delivery service in Herman v. Express Sixty-Minutes
Delivery Serv., Inc.102, to name just one example, that contractually required its
drivers to be on-call around the clock but otherwise allowed them to “set
their own hours and days of work and [ ] reject deliveries without
retaliation”103—a court might well take one look at the similarly permissive
quest-assignment engines of most MMOs and declare them just another
form of independent-contracting scheme.
Yet any court that took the time to understand MMOs as a whole
would likely recognize that the analogies don’t come so easily. Control as it is
exercised in a modern online game and control as it functions in a traditional
workplace are two very different things, and properly comparing them
requires some adjustment for the differences. Imagine, to begin with, a realworld workplace in which the proprietor controls the environment as totally
as an MMO developer controls its world: A delivery service, for example,
that can alter at whim not only the décor in the drivers’ break room but the
look, feel, and function of everything else they interact with: the cars they
drive, the roads they drive on, even the bodies they come to work in.
Imagine, next, that the real-world employer additionally has the MMO
developer’s power to fine tune not just the working environment but the
workers’ motivations—that it can dispense with the blunt incentive of regular
102
103
Herman v. Express Sixty-Minutes Delivery Serv., Inc., 161 F.3d 299 (1998).
Herman 161 F.3d at 303.
Page 42 of 68
DRAFT; please do not cite or circulate without author’s express permission
wages and instead deploy the proven behavioral-psychological methods of
operant conditioning, doling out loot and other rewards on a frequency curve
precisely calibrated to induce compulsive repetition of even the most
mindless tasks.104 Now, finally, consider whether, in so thoroughly controlled
a workplace, the ability of workers to “set their own hours and days of work”
or to “reject [assignments] without retaliation” can really be the same
persuasive indicium of independence that it was to the Herman court?
It cannot. Nor, once the full extent of the MMO’s control over its
players is conceded, do there remain among the factors and principles of the
control test any decisive obstacles to deeming MMO players employees of
the MMO and its developer.105 That the MMO developer’s control over
player performance is less direct than a conventional employer’s may be true,
but even if that necessarily made it less effective, control-test doctrine
explicitly abjures any inherent distinction between direct and indirect
control.106 A more qualified objection might be that MMO play fails more
robust versions of the control test, such as the variation often applied in
FLSA cases and sometimes characterized as a test of “economic reality.”107 In
this variant, courts may find employment of an individual by a hiring entity
See NICK YEE, THE PROTEUS PARADOX: HOW ONLINE GAMES AND VIRTUAL
WORLDS CHANGE US—AND HOW THEY DON’T 39–43 (2014) (discussing MMO
designers’ deliberate use of operant-conditioning reward schedules to structure the
leveling grind).
105 In principle, even gold farm employees can also be employed by the game
company in this sense, since the common-law test recognizes no inherent bar to a
worker’s having more than one employer. Martinez-Mendoza, 340 F.3d at 1208
(recognizing the possibility of joint employment under the FLSA).
106 Id. (“control may be either direct or indirect, taking into account the nature of the
work performed”).
107 See Zatz, supra note 93, at 871 n.50 (gathering “economic reality” cases).
104
Page 43 of 68
DRAFT; please do not cite or circulate without author’s express permission
“if, as a matter of economic reality, the individual is dependent on the
entity.”108 If dependence, in this sense, means dependence for survival, then
no unpaid game play could qualify. What appears to be meant by the word,
however, is simply a more ample notion of employer control, which in some
courts’ expressions of the concept seems, if anything, even likelier than the
standard test to find employment in the kind of control MMO players are
subject to.109
Objections even narrower are possible: That quest rewards and other
loot drops effectively pay players for projects completed rather than time
worked (an indicator of independent-contractor status under the control test).
Or that in using their own computers and Internet connections, players
effectively invest in their own work equipment (ditto).110 Yet as to the first
point, quests and other tasks in MMOs generally proceed to completion at
rates predictable enough that their rewards amount to piecework pay, which
courts repeatedly have found consistent with employment status.111 As to the
second, even the most avid MMO players use their computers and Internet
connections for more than just playing MMOs, and courts applying the
control test have only found equipment used exclusively for work to weigh
Martinez-Mendoza, 340 F.3d at 1208.
See Zheng v. Liberty Apparel Co., 355 F.3d 61, 72 (2d Cir. 2003) (calling the
“economic reality” variant a test of “functional control over workers even in the
absence of . . . formal control”).
110 Cherry makes this argument in applying the control test to virtual work in general.
Cherry, supra note 47, at [ ].
111 See Herman 161 F.3d at 306 (gathering cases holding pieceworkers of various
sorts to be employees).
108
109
Page 44 of 68
DRAFT; please do not cite or circulate without author’s express permission
against employee status.112 And even if both arguments proved valid, it would
take more than two minor adverse factors to tip the balance of an inquiry as
searchingly committed to the totality of the circumstances as the control
test.113
If control were the only issue that mattered in deciding whether
MMO play is or is not employment—as in most such decisions it is—then
the matter would stand decided. But one more question—typically reserved
for cases testing the outer limits of employment law—must be asked, and
that is, in effect, whether MMO play is or is not work.
B. Do Players Meet the Economic Test of Employment?
When the law must decide what does and does not fall under the
governance of the FLSA or some other employment-based statute, it
typically does so on one of two bases: It applies the control test or it applies a
statutory exemption. If the activity in question neither fails the control test
nor belongs to any category of employment explicitly exempted by the
statute, then the inquiry almost always ends there, with the activity
recognized conclusively as employment. In rare cases, however, the court
See id. at 304 (finding drivers’ personal use of delivery cars to weigh against their
being independent contractors). See also Brock v. Mr. W Fireworks, Inc., 814 F.2d
1042, 1051 (5th Cir. 1987) (discounting as evidence of independent contractor status
a worker’s work-related use of a home computer originally bought for school use).
113 See id. at 1054 (noting that in applying the control test “[i]t is erroneous to focus
on a single factor . . . and thereby fail to consider the entire circumstances of the
work relationship.” (quoting Brennan v. Partida 492 F.2d 707, 709 (5th Cir. 1974))).
112
Page 45 of 68
DRAFT; please do not cite or circulate without author’s express permission
may insist on a new stage of inquiry, aimed at a set of questions markedly
distinct from those of the control test.
What provokes the heightened scrutiny in these cases is an anomaly
that Zatz, in his work on the labor of prisoners, calls “paid nonmarket
work.”114 This includes not only prison labor but a range of other
phenomena—graduate student teaching, workfare programs, therapeutic jobs
for the disabled—in which people are paid for labor performed in
institutional contexts aimed primarily at rehabilitation or self-improvement.
As Zatz explains:
When these workers assert employment rights, they face fierce
resistance on the ground that their work lies outside of the labor
market. The dispositive legal question always is whether an
employment relationship exists. Courts determine the answer by
asking whether the relationship is economic in nature.115
Zatz traces the emergence of this new employment test through a
series of lawsuits by prisoners seeking minimum wages and other rights
under the FLSA for their labor while confined.116 For years, opponents of
these suits were powerless to stop them on the usual fronts: No statutory
exemptions for prison labor could be found in—or read into—the pages of
the FSLA, and the control test tended naturally to find imprisoned workers
Zatz, supra note 93, at 862.
Id.
116 Id. at 867–82.
114
115
Page 46 of 68
DRAFT; please do not cite or circulate without author’s express permission
to be pretty well controlled.117 Eventually, though, courts began to pick up on
a third approach first articulated by the D.C. District Court in Souder v.
Brennan,118 an FSLA wages and overtime suit brought by three involuntarily
committed mental patients. Under the circumstances, sufficient control for a
finding of employment could be presumed. The statute, likewise, held no
obviously applicable exemptions, nor could the court persuade itself that an
exemption was to be inferred. What interested the court instead was a
different sort of statutory hook: The FSLA’s definition of “employ” as “to
suffer or permit to work.”119 The phrasing suggested a new question by
which to gauge the plaintiffs’ employment status: Not whether their work
was under someone else’s control but whether it actually was work. Nor was
the answer obvious. Indeed, the defendants’ key claim was that the services
performed by patients—acting as “dishwashers, kitchen helpers, messengers,
and the like”120—were better understood as therapy than as work. But the
court ruled otherwise: The services were work, and they were work because,
whatever therapeutic benefit accrued to the patients providing them,
“consequential economic benefit” accrued to the institutions receiving
them.121 In short, there was now a new employment test: Regardless of who
controlled the laborer or how well, if there was not additionally an economic
relationship between them, there could be no employment relationship either.
Id. at 871–79.
Souder v. Brennan, 367 F. Supp. 808 (D.D.C. 1973).
119 Id. at 813 (quoting 29 U.S.C. § 203(g) (2006)).
120 Id.
121 Id. at 814 n.29.
117
118
Page 47 of 68
DRAFT; please do not cite or circulate without author’s express permission
According to Zatz’s reading of the prison labor cases, this new test
has evolved over the years into two mutually incompatible strains.122 The first,
adopted by the majority of courts, takes what Zatz calls an “exclusive market”
approach. Defining the required economic relationship as fundamentally a
market relationship, this version of the test recognizes employment only
where it has been entered into under the conditions of a classically free and
open market, chief among those conditions being liberty of contract,
mutually gainful exchange, and arms-length bargaining. As these do not often
typify relations between prison and prisoner, the practical effect of the
exclusive market test has been to invalidate whatever prisoner employment
claims it has been applied to.123
The other variant of the economic test, what Zatz calls the
“productive work” approach, just as reliably affirms employment status when
applied. Echoing Souder’s emphasis on the economic value of the work
performed—and indifference to the noneconomic values that may in part
have motivated the work—the productive-work test focuses almost
exclusively on the economic effects of the employment relationship,
particularly as they are felt by parties outside that relationship. The test’s
most full-throated expression perhaps has been a Ninth Circuit dissent
voicing, on the one hand, concern with the “pernicious competitive effect of
cheap [prison] labor”124 on outside labor markets and, on the other,
bafflement with the exclusive-market view of FLSA coverage as incompatible
Zatz, supra note 93, at 882–897.
Id. at 882.
124 Hale v. State of Ariz., 993 F.2d 1387, 1400 (9th Cir. 1993).
122
123
Page 48 of 68
DRAFT; please do not cite or circulate without author’s express permission
with the relationship between prison and prisoner.125 Crucial to such
determinations, Zatz notes, is the fungibility of inmate labor’s products—the
extent, that is, to which “[e]mployers of prison labor can substitute inmates
for other workers, and consumers can substitute products of inmate labor for
those produced by other means.”126 Thus, even where inmates’ work is
consumed solely by the incarcerating institution (as with in-house laundry or
food-preparation services), the fact that outside workers might otherwise
have been paid for it may lead the productive-work view to see economic
activity where the exclusive-market view would see activity better
characterized as rehabilitative.127
The question that concerns us here is which of these two views, if
either, would find that a player farming an MMO is doing work as the
economic test defines it. Certainly the productive-work view would have no
trouble finding consequential economic effects resulting from MMO farming,
whether it feeds directly into the RMT markets or not. The well-established
fungibility of virtual assets assures as much. From the perspective of a
Chinese gold farmer, after all, gold coins and other loot produced by unpaid
Western players enter the MMO’s virtual markets with a “pernicious
competitive effect” on his labor just as surely as his own products impinge
on the unpaid player’s efforts. Nor is it much more complicated to identify
the economic benefits flowing from the unpaid player to the putative
employer here, the game developer. As Miriam Cherry observes, in a cultural
Id. at 1403.
Zatz, supra note 93, at 893.
127 Id. at 895.
125
126
Page 49 of 68
DRAFT; please do not cite or circulate without author’s express permission
product as uniquely social as a virtual world, even basic farming activities can
add value to the product insofar as they make the virtual world more
interesting for other users.128 Thus, notes Cherry, “while performing a
[virtually productive] task may not directly benefit the company, it might
provide an indirect benefit.”129 And there seems to be little reason to assume,
in an age of billion-dollar social-media empires built on foundations of
virtual “likes” and “pokes,” that the indirect benefits Cherry imagines cannot
be precisely the sort of consequential benefit Souder’s economic test
contemplates.
Would the exclusive-market test’s additional requirements be any
harder for MMO play to meet? Possibly not. Though the core dynamics of
play may be different from the workings of markets, it would be hard for a
court to frame the two as mutually inimical in quite the same ways free
markets and prisons are. Indeed, as was previously discussed, existing law on
paid athletic play scarcely supports finding any such mutual antagonism at
all.130 Even if it did, the uniquely economic nature of MMO play itself might,
in this case at least, suggest relaxing any presumption of inherent
incompatibility between markets and play. As for the relationship between
player and developer—which is grounded, formally, in the terms of the
EULA and ToS and, functionally, in the exchange of consideration from
both sides—a court would likely find it more fundamentally contractual than
many real-world employment relationships.
Cherry, supra note 47, at 1102.
Id.
130 See supra Part II.B.2.
128
129
Page 50 of 68
DRAFT; please do not cite or circulate without author’s express permission
MMO play thus might well meet the explicit requirements of either
view of the economic test for employment. To say so, however, is not to end
the inquiry. As Zatz observes, there is a further requirement, implicit under
both views, that the employer-employee relationship not only produce
economic benefits but constitute a mutual exchange of economic benefits
between the parties.131 That is to say, the economic test has an exemption for
volunteers: a way for people to contribute economically productive services
to organizations without triggering wage requirements and other burdens of
employment law.132 Cherry likewise notes the exemption, citing the Supreme
Court’s holding in Walling v. Portland Terminal Co.133 that the FLSA does not
count as employees those who “without promise or expectation of
compensation, but solely for [their] personal purpose or pleasure, [work] in
activities carried on by other persons.”134
See Zatz, supra note 93, at 918–21.
This exemption isn’t fully tested in prison labor cases, since those generally
concern inmates paid money for their work, but Zatz finds it delineated in cases
where unpaid labor has given rise to claims of volunteerism. See id. at 921, citing
Graves v. Women’s Prof’l Rodeo Ass’n, 907 F.2d 71, 73 (8th Cir. 1990) (“[A]n
employer is someone who pays, directly or indirectly, wages or a salary or other
compensation to the person who provides services—that person being the
employee.”). Note that this mutuality requirement also creates an exemption for
interns and other trainees, who themselves gain economically valuable experience
through their work but do not in turn produce economic value for anyone else
through it. See Zatz, supra note 93, at 921, citing Walling v. Portland Terminal Co.,
330 U.S. 148, 153 (1947) (holding that railroad trainees were not employees because
“the railroads receive no ‘immediate advantage’ from any work done by the
trainees”).
133 Walling v. Portland Terminal Co., 330 U.S. 148 (1947).
134 Cherry, supra note 47, at 1100, citing Walling, 330 U.S. 148, 152. See also
Department of Labor, Volunteers (elaws - Fair Labor Standards Act Advisor),
http://www.dol.gov/elaws/esa/flsa/scope/ee16.asp (last visited Jan. 3, 2014)
(“[T]he Supreme Court has made it clear that the FLSA was not intended ‘to stamp
all persons as employees who without any express or implied compensation
agreement might work for their own advantage on the premises of another.’”), citing
Walling, 330 U.S. 148, 152.
131
132
Page 51 of 68
DRAFT; please do not cite or circulate without author’s express permission
Indeed, for Cherry, this exemption poses the principal obstacle to a
finding of employment for unpaid MMO farmers under the economic test.135
Yet here it’s important to note that, even assuming most MMO players play
primarily or even solely for their “personal purpose or pleasure,” it’s not clear
that the purpose or pleasure attained through farming can be meaningfully
distinguished from compensation. For one thing, as Cherry herself observes,
the readily ascertainable dollar value of virtual assets makes it easy to translate
the rewards of farming into an hourly wage.136 For another, even in the
absence of monetary compensation, courts may hold that contingent in-kind
benefits given to a worker can invalidate a claim of volunteerism.137 Finally,
and perhaps most decisively, authorities have further stated that under FLSA
workers may not volunteer their labor to for-profit enterprises.138 That MMO
developers are profit-seeking companies would not trigger this restriction, of
course, if players, like interns, could be presumed to provide no economic
benefit to their putative employers.139 But since we are presuming just the
opposite at this point, it is hard to see how the volunteer exemption could
stop an MMO player from being recognized as an employee where the
economic test alone has failed to.
Cherry, supra note 47, at 1096–1105 (presenting the control test and the volunteer
exemption as the two main hurdles to finding employment status for virtual workers
but considering only the latter in the MMO context).
136 Id. at 1103.
137 Tony & Susan Alamo Found. v. Sec’y of Labor, 471 U.S. 290, 301 (holding that
for the purposes of the volunteerism analysis, “in-kind benefits . . . expected in
exchange for [ ] services” are “another form of wages”).
138 See Department of Labor, supra note 134.
139 See supra note 132.
135
Page 52 of 68
DRAFT; please do not cite or circulate without author’s express permission
It remains, then, only to consider the consequences that might follow
from such recognition.
C. What Happens if Unpaid Farmers Are Employees?
Assimilating unpaid play into the employment-law regime might
seem to be the ultimate victory of commodification over play. This, in fact, is
precisely the concern that gives rise to calls by Castronova and others for
laws protecting virtual worlds from the reach of economic laws.140 Zatz,
moreover, provides some support for this concern in his criticisms of the
economic test as an analysis of economic reality. For Zatz, the problem with
both the productive-work and exclusive-market analyses is that they are
essentially binary: The exclusive-market view, on the one hand, refuses to
characterize as economic any relationship that is not purely so while the
productive-work view, on the other, renders wholly economic even those
that are only partly so.141 As applied to the productive-work view in particular,
this complaint broadly tracks the concerns of the anti-commodificationists
with market-based incursions on nonmarket spheres of life,142 flagging
especially the domino effects that may result from the application of
employment law to activities not previously governed by it.143 For example, a
relationship held to be employment under the FLSA’s economic-reality test
See supra text accompanying notes 87–89.
Zatz, supra note 93, at 895.
142 Id. at 923 (“The difficulty, then, is that when employment law intervenes in an
economic relationship, even with regard to its economic terms, it necessarily also
intervenes in the relationship's noneconomic aspects.”).
143 Id. at 923.
140
141
Page 53 of 68
DRAFT; please do not cite or circulate without author’s express permission
may become, as the result of imposed wage and hour conditions, cognizable
as employment under the NLRA, and so on. In practice, Zatz finds, an
inconsistent adoption of both the exclusive-market and productive-work
views by courts tends to provide a check on the “runaway tendencies” of
either.144 Yet there is less hope of such constraint where, as in the case of
unpaid MMO farming, the exclusive-market view may lead almost as
probably as the productive-market view to findings of employment. If
employment status for MMO players is plausible under any circumstances,
therefore, then those who fear the commodification of MMO play would
appear to be justified in opposing it at every turn.
Yet before they do, they might do well to remember that the
employment-law regime itself originated as a kind of bulwark against
commodification, a way of setting limits to the market’s control over our
working lives. To suggest that employment law could have a similarly
protective effect on our playing lives might be incongruous, but it’s not
absurd.
Consider employment law’s restrictions on work hours. As Cherry
notes, among the first elements of the FLSA that would kick in once players
become employees would be child labor laws.145 Among other things, these
laws place hard limits on the number of hours children can work. Would we
want to place such limits on the hours they can play? In the U.S. the question
is scarcely considered, yet in other jurisdictions the obsessive quality of
MMO and other online play—and related news reports of death and injury
144
145
Id. at 903–4.
Cherry, supra note 47, at 1100.
Page 54 of 68
DRAFT; please do not cite or circulate without author’s express permission
resulting from online-game “addiction”—has given rise to restrictions not
dissimilar to those of child-labor law.146 In China, for example, authorities
have required online game developers to implement an “Online Game AntiFatigue System” barring minors from the game for five hours after five hours
of continuous play.147 Korea, too, has proposed a curfew on underage online
play during a six-hour block of night.148 There is more than a hint of moral
panic in these responses, and players and game companies alike have either
contested or evaded them.149 Yet it is difficult to imagine what principled
objections to such restrictions could be made by those whose objections to
commodification of MMO play include complaints about the extra hours of
farming that paid farming forces all players to engage in.150
For that matter, why should they object to the application of FLSA
restrictions to adult play hours either? Though the FLSA puts no caps on the
number of hours adults can work, it does require overtime wages for work in
excess of a forty-hour week. China’s Online Game Anti-Fatigue System
already implements a somewhat analogous rule for underage players, reducing
by half any experience points or other quantifiable reward the game delivers
See generally Nachshon Goltz, “ESRB Warning: Use of Virtual Worlds By Children
May Result in Addiction and Blurring of Borders”: The Advisable Regulations in Light of
Foreseeable Damages, 11 U. Pitt. J. Tech. L. & Pol’y 2.
147 Id. at 43.
148 Christine Kim, South Korea To Put Curfew on Online Games for Kids, REUTERS, Apr
13 2010, http://in.reuters.com/article/2010/04/13/us-korea-onlinegamesidINTRE63C1AJ20100413 (last visited Apr 15, 2014).
See also Brian Ashcroft, Korea’s Silly “Shutdown Law” Might Be Unconstitutional,
KOTAKU, http://kotaku.com/5921659/koreas-silly-shutdown-law-might-beunconstitutional (last visited Apr 6, 2014).
149 See id. (describing game companies’ constitutional challenge to the proposed
Korean curfew) and Goltz, supra note 146, at 46 (describing underage players’ ability
to evade the Chinese play restrictions).
150 See Castronova, supra note 87; Hernandez, supra note 39.
146
Page 55 of 68
DRAFT; please do not cite or circulate without author’s express permission
after three hours of continuous play.151 Arguably, of course, increasing rewards
by half after forty hours of play might only incentivize the kind of power
gaming that the anti-commodificationists complain about. But then again, it
might have the more targeted effect of incentivizing only those who already
find the farming grind a peculiar sort of fun, while allowing those for whom
it is a more burdensome means to a quantifiable end to reach that end sooner
once the time investment exceeds that of a normal work week.
The FLSA’s minimum wage requirements might serve as a similarly
counterintuitive check on the effects of commodification. They would be
trickier to apply, of course, since they would require game companies both to
track the real-money value of in-game assets and, whenever the rate at which
that value became available to players falls below $7.25 an hour, increase that
rate without either restricting or expanding the supply of assets so much that
the increase is negated. As a practical matter this would be burdensome but,
given that at least one MMO keeps an accredited economist on staff to deal
with its virtual money supply, presumably not unreasonable.152 The more
philosophical question is whether MMOs would be sustainable as businesses
if required to maintain their reward rates at minimum-wage levels.
Considering that Castronova found an effective hourly wage of $3.42 in the
leading MMO of 2001, this is no idle question.153 But it’s worth noting, as
well, that it is essentially the same question that has been asked about
Goltz, supra note 146, at 43.
See Neal Ungerleider, Meet the Alan Greenspan of Virtual Currency in “EVE Online”,
FAST COMPANY (Jan. 6, 2014), http://www.fastcompany.com/3024392/meet-thealan-greenspan-of-virtual-currency-in-eve-online.
153 See supra note 14.
151
152
Page 56 of 68
DRAFT; please do not cite or circulate without author’s express permission
minimum-wage requirements for as long as they have existed. And assuming,
as has arguably been the case with most extensions of the minimum wage,
that this one does not significantly harm the businesses to which it applies, it
is again difficult to think why its effects would not be welcomed by people
critical of MMO play’s convergence toward full-time work.
But the application of standardized hours caps and wage floors to the
unique context of MMOs need not be the only effect of finding employment
in the midst of that context. As unionized workers, employees are
empowered—in ways that they are not in purely contractual work
relationships—to negotiate terms tailored to the specific conditions of their
workplaces. If recognized as employees under the NLRA, MMO players
could form unions and find themselves similarly empowered. In some ways,
this would not be as drastic a transformation as it might seem. MMO players
already make their interests known to and felt by MMO developers in a
variety of coordinated ways, including participation in player discussion
forums provided by developers, attendance at conferences hosted by game
companies, and formation of teamlike “guilds” composed of dozens and
even hundreds of players. It is not entirely clear what unionization would add
to this array of coordinating mechanisms. Whether the resulting workstoppage rights could be construed to give force to threats of collective
boycott is hard to say. But at the very least, implementation of NLRAsanctioned, game-wide player organizations would put players in a legal
relationship with one another, thus countering what could be considered a
key commodifaction effect at work in MMOs: The consignment of playerPage 57 of 68
DRAFT; please do not cite or circulate without author’s express permission
developer relations to the strictly bilateral model of consumer contracts,
complained about by Fairfield in his discussion of the Hernandez case.154
This is not to say that any of these developments is likely. Even if a
finding of employment status for MMO play is consistent with existing case
law, most courts will no doubt be reluctant to extend employment law so
sweepingly to an activity so ostensibly dedicated to the principle of play. But
working through the likely consequences of such an extension suggests,
ultimately, that it would not necessarily usher in the “contagion of reality”
that Caillois feared and Castronova still complains about. In fact, and to the
contrary, that same contagion might be checked through employment laws
more effectively than through the contractual status quo.
IV. ARE PAID FARMERS PLAYERS?
A. Are Gold Farmers Actually Playing the Game?
The law, we have seen, has little difficulty accepting that an activity
can be at once work and play. But it also has no trouble drawing lines
between what is play and what is not.155 In many cases, for example, courts
determine whether an activity does or does not conform to the rules of a
game and thus, in effect, whether or not someone engaged in that activity is
actually playing the game.156 Whether that activity is or is not play remains, of
See supra note 42.
See supra notes 75–77 and accompanying text.
156 See supra notes 78–85 and accompanying text.
154
155
Page 58 of 68
DRAFT; please do not cite or circulate without author’s express permission
course, a logically separate question.157 But in practice, deciding the first
question tends to look a lot like deciding the second.158 To find that someone
in some way is not playing a particular game is, for most courts and most
cases, to find that they are not playing at all.
Whether Chinese gold farmers are playing at all is, as we saw, not
easily determined just by looking at their attitudes toward and day-to-day
experience of the games they work in.159 But it isn’t clear that just by looking
instead at the rules of those games we make the problem any easier. To be
sure, there is no ambiguity in the MMO companies’ bans on RMT and gold
farming, and it would seem to be within reason for a court to consider any
such ban a game rule. Unlike the casino in Uston, after all, an MMO
developer is fully authorized to set its own rules for its own games. Why then
shouldn’t any contractual provision by which a developer governs players’
behavior be deemed a rule defining the game they play?
Whether the one determination follows from the other depends to some extent
on one’s choice of play theory. A number of games scholars in recent years have
written on the subject of “transgressive play,” observing that individuals who violate
the rules of a game may see themselves as—and may in fact be—playing by the
rules of a different game altogether. See MIA CONSALVO, CHEATING: GAINING
ADVANTAGE IN VIDEOGAMES (2009); Espen Aarseth, I fought the law: Transgressive
play and the implied player, SITUATED PLAY. PROC. DIGRA 24–28 (2007). In classic
theories of play, however, the behavior of the cheater, the spoilsport, and others
who fail to conform to a game’s rules is viewed, in varying degrees, as a negation of
true play. See HUIZINGA, supra note 61; CAILLOIS, supra note 63; SUITS supra note 65.
158 This is true, for example, in cases where in-game behavior falls so far outside the
bounds set by the rules that it becomes actionable as tort. See Hackbart 601 F.2d at
520 (unnecessary roughness tort case). It is also true where the behavior has so little
relevance to the game’s rules that it becomes subject to legal protections not usually
thought to govern play. See Martin, 532 U.S. at 683 (holding that a golfer’s riding
rather than walking between shots did not affect the nature of the game and was
therefore covered by the Americans with Disabilities Act’s public accommodation
provisions); Uston 89 N.J. at 168, 173 (holding that a player’s counting cards was
irrelevant to the actual rules of casino blackjack and therefore covered not by those
rules but by New Jersey’s public accommodation laws).
159 See supra text accompanying note 25.
157
Page 59 of 68
DRAFT; please do not cite or circulate without author’s express permission
Yet the same question can be asked of the PGA and the walking
requirement it sought to enforce as a game rule in Martin. And the Supreme
Court’s holding there suggests that a game owner’s proprietary authority
cannot always be enough, on its own, to turn a contractual requirement into
a game rule. In Martin, the court also considered both the overall design and
historical evolution of golf before finally deciding that the walking rule was
not among the game’s “essential attribute[s].”160 MMOs, and RMT-related
issues in particular, would seem to invite a similarly searching evaluation of
game-rule claims. The multiplicity of play styles found in an MMO, the
diversity of feelings about RMT, and above all perhaps the intensely social
and essentially economic nature of MMO play—all these urge founding the
inquiry on an assumption, as in Martin, that the game’s rules are not simply
what the game’s owner says they are.
The inquiry itself would necessarily be a fact-intensive one, possibly
verging on the ethnographic, and predicting the outcome for any given
MMO or rule would likewise verge on the impossible. What cannot be ruled
out in any case is that courts, asked to decide whether an MMO’s RMT ban
is an “essential attribute of the game itself,” will hold that it is not. Thus, just
as in post-Uston New Jersey casinos the law dictates that a professional card
counter is as legitimate a player of blackjack as any other,161 so too might
there one day be MMOs in which the range of legitimate play styles must, as
a matter of law, include that of the professional farmer.
Martin, 532 U.S. at 685.
See Uston, 89 N.J. at 169 (“[A casino] has no right to exclude [a card counter] on
grounds that he successfully plays the game under existing rules.”).
160
161
Page 60 of 68
DRAFT; please do not cite or circulate without author’s express permission
B. What Happens if Paid Farmers Are Players (and if They Are Not)?
The legal consequences of finding gold farmers to be players would
likely be less direct and less sweeping than those of finding unpaid farmers to
be employees. But they could nonetheless determine significant legal
outcomes.
Uston provides the obvious template for a scenario in which the
question of gold farmers’ status as players proves decisive. Indeed, were the
New Jersey Supreme Court to hear a gold farmer’s challenge to an MMO
company’s RMT ban—like the one brought in Blacksnow, for example162—it
might well find Uston to control. In Uston, the court held that a property
owner’s common law right to exclude, which in most states is limited only by
statutory exceptions, in New Jersey “is substantially limited by a competing
common law right of reasonable access to public places.”163 That an online
game might be deemed a public place for the purpose of the rule is hardly
inconceivable.164 Once reached, such a finding would leave no very clear
remaining grounds for distinction. For example, though the Uston court
suggested blackjack professionals might lose their right of reasonable access
See supra note 32 and accompanying text.
Uston, 89 N.J. at 168.
164 See, e.g., Nat’l Ass’n of the Deaf v. Netflix, Inc., 869 F. Supp. 2d 196, 200–2 (D.
Mass. 2012) (holding that Netflix’s streaming video site was a public
accommodation under Americans with Disabilities Act provisions covering “place[s]
of exhibition and entertainment” or “place[s] of recreation”). But see Cullen v.
Netflix, Inc., 880 F. Supp. 2d 1017, 1024 (N.D. Cal. 2012) (holding that the Netflix
site cannot be a public accommodation under Ninth Circuit precedent because it is
not “an actual physical place”).
162
163
Page 61 of 68
DRAFT; please do not cite or circulate without author’s express permission
if their methods were inherently “disorderly” or otherwise disruptive of a
casino’s “essential operations”165, the difficulties of showing just how RMT
harms other players would complicate any attempt to bring that exception to
bear on gold farmers.166 Likewise, although the state’s exclusive authority to
promulgate casino blackjack rules in New Jersey has no equivalent in the
MMO context, that difference, as we have noted, does not necessarily vest
exclusive rulemaking authority in MMO companies either. Aside from the
Martin court’s refusal to defer to the game operator’s formulation of the
game rules, copyright’s doctrine that the game operator cannot own whatever
rules it formulates167 further undermines whatever claim to arbitrary authority
an MMO developer might want to make.
Again, none of this is to suggest that such a case is likely to occur.168
But the possibility illustrates how finding gold farmers to be players might
affect key procedural issues in particular. Just as in Uston a finding that the
casino’s ban on card counting was a game rule would presumably have
stopped the court from reaching the loftier questions of property law on
which it ultimately decided the case, so too a finding that an MMO’s ban on
Uston, 89 N.J. at 173.
See supra notes 27–30 and accompanying text. The parallels with blackjack, again,
run deep. Because blackjack players play only against the house, the game, like
MMOs, is not zero-sum competitive and should therefore leave players relatively
indifferent to the success or failure of others in the same game. Yet even so, players
will often complain when tablemates make rookie plays that turn out to favor the
dealer and thus hurt the whole table, even though a smart play could just as likely
have had the same effect. See John Grochowski, Blackjack Etiquette and Strategy,
HOWSTUFFWORKS, http://entertainment.howstuffworks.com/how-to-playblackjack.htm (last visited Apr 14, 2014) (giving novice players tips on how to keep
their subcompetent play from drawing the ire of the more experienced).
167 See supra note 85 and accompanying text.
168 If nothing else, the typically global reach of online games would make it easy for
an MMO company to remove the suit to a venue where New Jersey’s unusual public
access doctrine does not apply.
165
166
Page 62 of 68
DRAFT; please do not cite or circulate without author’s express permission
RMT is a game rule, even if it had to be supported with messy, complicated
evidence of player practices and game-design imperatives, be worthwhile for
the MMO company to pursue if only as a means of narrowing the issues in
play. A similar effort might arguably be advised for the plaintiff in any future
suit replaying the strategy in Hernandez. Making the case that an MMO’s RMT
ban is a rule of the game could, for example, help block any challenges to the
plaintiff’s otherwise fragile standing as a third-party beneficiary of the
contract between MMO and gold farmer. As long as the promise to refrain
from RMT is just a provision of that contract, the court may doubt that the
parties intended it to benefit all other players without plain language stating
that intention in the writing. But if the provision is deemed additionally to
constitute a rule—or in Martin’s terms “an essential attribute of the game
itself”—then the intention is much more easily inferred. The “game itself” is
after all precisely and by definition what the company provides to every
player, so any provision aimed at protecting the essence of the game must
necessarily be understood to benefit all players. Indeed, having established
that basic a harm to her interests from the practice of RMT, the plaintiff
might even be able to win declaratory and injunctive remedies, at least,
without having to wade into the difficulty of proving more concrete harms.
Finally, too, note that even though the question of players’ employee
status may ultimately be more consequential than that of farmers’ player
status, the latter could turn out to constrain the former in one important way:
The reach of any finding that players are employees of the game company
might be decided in significant part by determining which users of the
Page 63 of 68
DRAFT; please do not cite or circulate without author’s express permission
company’s product, in fact, are players of its game. This might be so if for no
other reason than that a court finds the determination to be a reliable proxy
for deciding whether the farmer’s activity in the game creates a net economic
benefit to the company and thus the basis for an employment relationship.
In that case, if the ban on RMT is not a rule of the game, then gold farmers
are players and therefore endowed, presumably, with whatever employee
rights all players enjoy. If, on the other hand, the ban is a rule, then the only
companies that should ever have to treat gold farmers as employees will be
the ones that pay them their real-money wages.
CONCLUSION
Let me now clarify: This is not an article about the future. Gone are
the days when it was even plausible to predict that virtual worlds would one
day be the universal interface through which we access the oceanic volumes
of data already surrounding and shaping us. Compared to the social media
and mobile apps whose growth began to eclipse theirs several years ago,
virtual worlds are now more or less a backwater. If there are reasons to spend
time thinking through potential reconfigurations of the laws of work and play
within the context of a single subrealm of that backwater, they do not rest on
any likelihood that the employment law of MMOs is poised to become a
topic of particular social and/or economic urgency. The point, rather, is to
glance, from what may be a particularly enlightening angle, at a question
much broader and for quite a long time now more urgent: The momentous
Page 64 of 68
DRAFT; please do not cite or circulate without author’s express permission
instability of the boundaries by which work, in a given age and culture, is
defined.
Of the factors that contribute to that instability, the present exercise
shines light on two in particular. The first is technology and its evolutions.
This, of course, is a subject already much discussed within the
commodification debate. From its beginnings in Marx’s analysis of alienated
factory labor, in fact, commodification theory has been sensitive to the role
of technology in bringing core human attributes within the ambit of market
exchange. More recently, in turning its attention to organ trafficking, in vitro
fertilization surrogacy, and other increasingly intimate forms of marketmediated self-alienation, commodification theory has remained alert to
emerging practices borne of and conditioned by advances in industrial and
postindustrial technology.
The rise of the Internet has been no less productive of new problems
in commodification theory. As Radin has noted, merely by creating new and
relatively unfettered channels for market exchange, the Internet has vastly
expanded the markets for such already contested commodities as human
eggs and Nazi paraphernalia.169 But it has also created markets in new forms
of commodified and semi-commodified labor, at or beyond the limits of
what is recognizably compensated work: the hypercasual “microlabor”
brokered by sites like Amazon Mechanical Turk and Task Rabbit; the
ubiquitous triangle trade in “user generated content” between users, service
providers, and advertisers on sites like Facebook and Twitter and Google
Margaret Jane Radin and Madhavi Sunder, Introduction: The Subject and Object of
Commodification, in RETHINKING COMMODIFICATION, supra note 49 at 8–9.
169
Page 65 of 68
DRAFT; please do not cite or circulate without author’s express permission
Search; even to some extent the open source software production system,
built on volunteer work and celebrated by some as the antithesis of
commodified labor but productive of massive economic benefits to the
software and Internet industries and of complex forms of compensation to
the producers.170
Scholars both inside and outside of legal academia have begun to
grapple with these new forms and to gauge the extent of their challenge to
existing social and regulatory regimes governing labor.171 The labor of MMO
farmers, both paid and unpaid, is yet another of these boundary-bending
forms of digital work, and its similarities to all the rest is part of what makes
it a potentially illuminating point of comparison. But even more illuminating
is what sets it apart: The degree to which it is marked as play. This is useful
because play, while not as markedly identifiable in other kinds of digital labor,
pervades them. Under the guise of leisure, creativity, amusement, passion,
play is a key motivating element across the landscape of online production.
To fully understand how these new forms of labor function, then, it may be
170
See generally Eli M. Noam, The Economics of User Generated Content and Peer-to-Peer:
The Commons as the Enabler of Commerce, in PEER-TO-PEER VIDEO 3–13 (Eli M. Noam
& Lorenzo Maria Pupillo eds., 2008.
171 For legal scholars, see, e.g., Cherry, A Taxonomy of Virtual Work, supra note 47;
Jonathan Zittrain, Ubiquitous Human Computing, 366 PHILOSOPHICAL
TRANSACTIONS OF THE ROYAL SOCIETY A: MATHEMATICAL, PHYSICAL AND
ENGINEERING SCIENCES 3813–3821 (2008). Dan Hunter & F. Gregory Lastowka,
Amateur-To-Amateur, 46 WM. & MARY L. REV. 951 (2004). For nonlegal scholars, see,
e.g., DIGITAL LABOR: THE INTERNET AS PLAYGROUND AND FACTORY (Trebor
Scholz, ed., 2012), Tiziana Terranova, Free Labor: Producing Culture for the Digital
Economy, 18 SOCIAL TEXT 33–58 (2000).
Page 66 of 68
DRAFT; please do not cite or circulate without author’s express permission
not only useful but in some sense necessary to begin the inquiry with a focus
on the most clearly ludic of them. To explore the limits of MMO farming’s
capacity to bend to existing legal categories of work and play is thus, to an
extent, to consider how those categories interact in online settings
generally—and to prepare, perhaps, new strategies, both legal and nonlegal,
for handling the new kinds of exploitation and opportunity so peculiarly
playful a work environment presents.
But if the peculiarities of the online setting itself are part of what
makes work so Protean a category there, the other and probably more
important part is that work happens to be a Protean category to begin with.
For Zatz, recall, the interest in studying how courts decide what is work and
what is not is not that it brings us any closer to actually understanding what
work is; what it shows, rather, is how courts, through their decisions, are
continuously helping to construct and reconstruct what we recognize as
work.172 Yet if directly studying those decisions is one way to get close to that
constructive process, trying to imagine how those decisions might be adapted
to some new set of circumstances—in effect, to imagine the next iteration of
the process oneself—gets us arguably even closer.
And that, in the end, is probably the real value of this exercise. As a
culture and society, we may be closer now than at any time in the last three
centuries to accepting that work and play may not, in fact, be mutually
exclusive categories. But even so, the thought of assimilating the
unadulterated play of the unpaid MMO player to the laws of employment—
172
See supra notes 57–58 and accompanying text.
Page 67 of 68
DRAFT; please do not cite or circulate without author’s express permission
or, likewise, of rendering the wage-bought play of the professional gold
farmer legally equivalent to any other’s play—still has a whiff of paradox
about it. To proceed nonetheless to think through the concrete steps it might
take to reach either of those paradoxical ends, and thereupon to learn that
both in fact lie just a few short leaps of reasoning away from existing case law,
is to understand at last just how malleable the concepts of work and play can
be.
Page 68 of 68